Upskilling Infrastructure for Data Scientists
Become the most versatile
data scientist on your team.
Self-paced courses, monthly masterclasses, targeted workshops. Built by 20+ practitioners around the world using the commercial tools companies actually pay for. No surface-level tutorials — real projects, real code, real tools.
What Membership Unlocks
Three pillars of continuous upskilling
~/athlyticz/llm-workflow.R
R
Pillar 01 · Self-Paced
Deep, project-based courses
Production-grade courses covering the full stack: tidyverse, Shiny, Python, Stan/PyMC, machine learning, LLMs in R, biomechanics, plus the new Rapid App Prototyping / Agent Series (104 lessons launching this summer). Hundreds more project-based lessons over the next year — taught through sports data, directly applicable to finance, pharma, healthcare, and any industry.
751+
Lessons Live
104
Coming July
1000s
By 2027
July 2026
● LAUNCHING
S
M
T
W
T
F
S
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
10+ sessions each month120+/year
Pillar 02 · Live & Recurring
120+ masterclasses per year
Live sessions every month starting this summer. Deep dives with world-class instructors, Q&A, working sessions. Attend live or watch the replay. Every session ships with enterprise Posit Cloud VMs so you code along in the same environment as the instructor.
10/mo
Live Sessions
Posit
Cloud VMs
100%
Recorded
Causal Inference
D3.js for Shiny
APIs & Webscraping
AI Prototyping
Interview Prep
Quarto
Causal Inference
D3.js for Shiny
APIs & Webscraping
AI Prototyping
Interview Prep
Quarto
Databricks
SQL Mastery
Finance / Quant
Git Workflows
Mobile Apps
Sports Analytics
Databricks
SQL Mastery
Finance / Quant
Git Workflows
Mobile Apps
Sports Analytics
Pillar 03 · Workshops
Targeted niche workshops
Focused workshops on specific problems — causal inference, APIs, AI prototyping, interview prep, Databricks, and more. Taught by practitioners from 6 countries who are shipping this work in production right now. Purchased separately; members receive discounted pricing.
20+
Instructors
6
Countries
19
Focus Areas
Commercial tools teams actually use
R
Python
Stan
Shiny
tidymodels
PyMC
ellmer
ragnar
Databricks
Quarto
All of this goes live in July 2026.
AthlyticZ Membership launches this summer with every course and every masterclass, plus enterprise Posit Cloud VMs so you code in the same environment as the instructors. Workshops available separately. One subscription. Everything below. Plus everything shipping next.
The Self-Paced Catalog
Every lesson. Every module.
Nothing hidden.
Click any course to expand. Every lesson is described so you know exactly what you're getting. Search titles. Filter by language or level. Don't miss the Rapid App Prototyping / Agent Series slotted in after the flagship courses — 104 new lessons launching July 2026.
01
Becoming a BayeZian I
The rigorous introduction to Bayesian thinking. Stan, priors, posteriors, and the intuition behind every inference.
Taught by
Dr. Scott Spencer
Dr. Scott Spencer
§01
1. Introducing Bayesian Analysis for Sports
Introduction
Kick off the course with context on why Bayesian inference matters and what you'll be able to build by the end.
Course Topics
A walkthrough of the modules you'll complete and how each layer of the course builds on the last.
§02
2. Exploring Uncertainty & Variation
Uncertainty & Variation
Understand how uncertainty shows up in real data and why every serious model needs to account for it.
Example — 100 Meter Olympic Sprint
A worked example that demonstrates the technique on a concrete problem.
Visualizing the Example Data
A worked example that demonstrates the technique on a concrete problem.
Quiz: Exploring Uncertainty & Variation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
3. Introducing Probability, Random Variables & Distributions
Probability, Random Variables & Distributions
Build the probability foundation every Bayesian model rests on — from random variables to joint distributions.
Random Variables
Work with the building blocks of probabilistic modeling and understand how to represent uncertain quantities.
Discrete Bernoulli & Binomial Distributions
Model binary and count outcomes — the distributions behind wins/losses, completions, and shot conversions.
Poisson Distribution
Model counts of events in fixed intervals — goals scored, pitches thrown, shot volume — the workhorse count distribution.
Counts Approach Normal Distribution
Why the normal shows up everywhere, when to trust it, and when a heavier-tailed distribution is a better fit.
Continuous Distributions & the Uniform
Understand the key families of distributions you'll reach for and when each one fits the data.
Continuous Uniform Distribution
The most information-free prior and when it's actually the right choice for your model.
Beta Distribution
The distribution for probabilities themselves — and why it's the prior of choice for percentages and rates.
Normal Distribution
Why the normal shows up everywhere, when to trust it, and when a heavier-tailed distribution is a better fit.
Summary Statistics
Compute and interpret the core descriptives that ground every analysis before modeling.
Joint Distributions
Model how two or more random variables move together — the foundation of multivariate inference.
Marginal Distributions
Extract one-variable insights from multi-variable models — essential for interpreting Bayesian posteriors.
Conditional Distributions
Reason about what the data tells us about one quantity given another — the engine of Bayesian updating.
Independence between Variables
Know when variables actually move independently and when that assumption breaks your model.
Getting to Bayes Rule
The mechanical heart of Bayesian inference — how prior beliefs update in the face of new evidence.
Quiz: Probability, Random Variables, and Distributions
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
4. Priors, Likelihoods, and Posteriors
Priors, Likelihoods, and Posteriors
The three ingredients of every Bayesian model, and how they combine to produce calibrated uncertainty.
Likelihoods
Specify how your data could have been generated under the model — the bridge from theory to evidence.
Normalizing Constant
Understand the denominator in Bayes' rule, why it's often intractable, and why MCMC exists to sidestep it.
Conjugate Priors
Conjugate priors — the mathematically convenient prior families that lead to closed-form posteriors.
Quiz: Priors, Likelihoods, and Posteriors
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
5. Simulating Distributions in R
Simulating Distributions in R Intro
Generate synthetic data from your model to validate assumptions and test analyses.
Transforming Random Numbers to Distributions
Build forms that validate input and guide users to correct answers.
Discrete Distributions
Understand the key families of distributions you'll reach for and when each one fits the data.
Quiz: Simulating Distributions
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
6. Random Variable Code Objects
Representing Distributions with a Random Variable Code Object
Work through this lesson to build fluency with Bayesian inference, applied to sports analytics problems.
§07
7. Simulations and Models in Stan
Simulations and Models in Stan Intro
Generate synthetic data from the model to test assumptions and validate downstream work.
Stan Documentation
Write documentation that helps your future self and collaborators.
Toy Stan Example, Simulating Values
A worked example that demonstrates the technique on a concrete problem.
Second Example, Beta Binomial
A worked example that demonstrates the technique on a concrete problem.
Quiz: Simulations and Models in Stan
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§08
8. Posterior Simulation, Example with Grid Approximation
Grid Approximation Example
A worked example that demonstrates the technique on a concrete problem.
§09
9. Approximate Posteriors with MH and HMC
Approximate posteriors with MH and HMC Intro
Hamiltonian Monte Carlo — the sampler behind Stan that makes high-dimensional problems tractable.
Hamiltonian Monte Carlo
Use random simulation to solve problems that are hard to approach analytically.
§10
10. A Language for Describing Models
A Language for Describing Models Intro
Work through this lesson to build fluency with Bayesian inference, applied to sports analytics problems.
§11
11. Simple Normal Regression
Simple Normal Regression Intro
Work through this lesson to build fluency with Bayesian inference, applied to sports analytics problems.
Coding a Normal Regression Model
Build this model — the applied practice of probabilistic modeling.
Compiling & Fitting the Model
Estimate your model's parameters and diagnose whether the fit is good enough to trust.
Checking HMC Diagnostics
Hamiltonian Monte Carlo — the sampler behind Stan that makes high-dimensional problems tractable.
Reviewing the Model Parameters
Inspect the fitted model's parameters and validate that it behaves as intended.
Quiz: Simple Normal Regression
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§12
12. cmdstanr Model Object, Helper Functions, Model Evaluation
cmdstanr Model Object, Helper Functions, Model Evaluation Intro
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Posterior Predictive Checks, Three (3) Approaches
Validate the model by simulating data from it and comparing to observed reality. The essential sanity check.
First and a Half Approach
Your first attempt at the problem — a deliberate baseline before adding sophistication.
Third Approach
An alternative method for tackling the problem — compare approaches and understand the tradeoffs.
Model Comparison, ELPD, and LOOCV
Compare competing models rigorously and pick the one that best explains the data.
Quiz: Cmdstanr Model Objects
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§13
13. Extending Normal Regression
Extending Normal Regression Intro
Extend the default behavior to meet your specific needs.
Categorical Predictors
Use a trained model to generate predictions on new data.
Multiple Predictors
Use a trained model to generate predictions on new data.
Quiz: Extending Normal Regression
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§14
14. Generalized Linear Models, A Conceptual Introduction
Generalized Linear Models Intro
Generalized linear models — extend linear regression to counts, probabilities, and other non-normal outcomes.
Logit Link Function
Write and use functions — the primary way to structure reusable, testable code.
Log Link Function
Write and use functions — the primary way to structure reusable, testable code.
Quiz: Generalized Linear Models
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§15
15. GLMs, Modeling Integer or Count Outcomes
GLMs
Generalized linear models — the statistical framework that spans logistic, Poisson, and other non-Gaussian responses.
Binomially-Distributed Count Outcomes
Work through this lesson to build fluency with Bayesian inference, applied to sports analytics problems.
Example Model 1 in Basketball
A worked example that demonstrates the technique on a concrete problem.
Example Model 2 in Basketball
A worked example that demonstrates the technique on a concrete problem.
Example Model 3 in Basketball
A worked example that demonstrates the technique on a concrete problem.
Poisson-Distributed Count Outcomes
Model count outcomes using the Poisson — appropriate when events occur independently at a constant rate.
Example Two Using Soccer Data
A worked example that demonstrates the technique on a concrete problem.
Example Three Using Soccer Data
A worked example that demonstrates the technique on a concrete problem.
Example Four Using Soccer Data
A worked example that demonstrates the technique on a concrete problem.
Quiz: Generalized Linear Models (Part 2)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§16
16. More GLMs, Modeling Categorical Outcomes
More GLMs: Modeling Categorical Outcomes
Dirichlet priors for categorical distributions — the Bayesian way to handle multi-class probabilities.
First Categorical Model
Model categorical outcomes where the predictor levels drive different response patterns.
Second Categorical Model
Model categorical outcomes where the predictor levels drive different response patterns.
Third Categorical Model
Model categorical outcomes where the predictor levels drive different response patterns.
§17
17. Hierarchical Models, an Introduction
Hierarchical Models Intro
Model nested data structures — players within teams, shots within games — the most powerful Bayesian pattern you'll learn.
Parameters Sharing Information
Build forms that validate input and guide users to correct answers.
Diagnostics and Reparameterization
Rewrite models to sample more efficiently — the technique that makes hard problems actually fit.
Quiz: Hierarchical Models
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§18
18. Workflow Recap
Workflow Recap
Bundle preprocessing and modeling into a single reproducible object.
§19
19. Case Study: Soccer and Expected Goals
Case Study Intro
Work through this lesson to build fluency with Bayesian inference, applied to sports analytics problems.
Visually Exploring the Pitch Data
Analyze pitch-level baseball data — the richest domain for practicing statistical modeling.
Modeling Goals as Bernoulli
Model goal-scoring as a probabilistic process — the foundation of soccer and hockey analytics.
Expanding the Model
Add complexity to the model — richer structure that captures more of the reality in your data.
Adding Angle Between the Goal Posts
Incorporate geometric structure into the model — angles, distances, and shapes as predictors.
Adding Predictor for Body Part
Use a trained model to generate predictions on new data.
Modeling Correlation Between Predictors
Measure relationships between variables and understand when correlation reflects real structure.
Adding Hierarchical Information to the Model
Model nested data structures — players within teams, shots within games — the most powerful Bayesian pattern you'll learn.
Reparameterizing the Model
Build this model — the applied practice of probabilistic modeling.
Using the Model Estimates for Decision Making
Translate model outputs into actionable choices — the step most courses skip.
§20
20. Next Steps
Next Steps
Your roadmap for going further — the resources, packages, and patterns worth exploring next.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
02
Becoming a BayeZian II
Hierarchical models, Gaussian processes, custom likelihoods. The full practitioner toolkit for real-world Bayesian work.
Taught by
Dr. Scott Spencer
Dr. Scott Spencer
§01
1. Introduction
Welcome to the Course
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
High-Level Review of First Course Topics
Consolidate what you've covered so far and confirm you're ready for what comes next.
Roadmap of this Course
See how the modules connect and where each skill plugs into the bigger system you'll build.
§02
2. Workflow
Workflow Introduction
Bundle preprocessing and modeling into a single reproducible object.
Before Fitting a Model
Estimate your model's parameters and diagnose whether the fit is good enough to trust.
Fitting a Model and Working with Simulations
Estimate your model's parameters and diagnose whether the fit is good enough to trust.
Evaluating and Using the Fitted Model
Measure whether your LLM actually works — eval datasets, scoring, and statistical comparison of models.
Understanding and Comparing Multiple Models
Compare candidate models side-by-side so your choice rests on evidence, not intuition.
§03
3. Mixture Models
Mixture Models Introduction
Mixture models — represent populations with multiple subgroups using weighted distributions.
Overdispersion
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Exploring Baseball Scores as Poisson Distributed
Explore the data systematically — the step that grounds every subsequent modeling choice.
Negative Binomial as Mixture of Poisson Distributions
Model counts of events in fixed intervals — goals scored, pitches thrown, shot volume — the workhorse count distribution.
Baseball Scores as Mixture of Poisson Distributions
Model counts of events in fixed intervals — goals scored, pitches thrown, shot volume — the workhorse count distribution.
Zero-Inflation Processes and Hurdle Models
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Modeling Three Point Attempts with Poisson and NB Models
Count regression with Poisson and its overdispersion-friendly cousin, the negative binomial.
Zero-Inflated Poisson
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Zero-Inflated Negative Binomial
Count regression with Poisson and its overdispersion-friendly cousin, the negative binomial.
§04
4. Rating and Ranking Models
Rating and Ranking Models Introduction
Rate or rank entities probabilistically — Elo, Bradley-Terry, and Bayesian ranking approaches.
Pairwise Comparisons
Compare multiple groups at once while controlling for false positives across comparisons.
Comparing Among Items in a Set
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Extended Ranking Models
Rate or rank entities probabilistically — Elo, Bradley-Terry, and Bayesian ranking approaches.
Fitting Plackett-Luce with stan
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Estimating Expectation of Position
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Ordinal Regression
Model ordered outcomes — ratings, rankings, severity scales — where the levels have a natural order.
General Principles of Ordinal Regression
Model ordered outcomes — ratings, rankings, severity scales — where the levels have a natural order.
Ordinal Regression in R and Stan
Model ordered outcomes — ratings, rankings, severity scales — where the levels have a natural order.
Simulating Scout Scores in Stan
Model scout ratings and subjective assessments as structured probabilistic quantities.
§05
5. (A Bit More) Advanced Hierarchical Models
Advanced Hierarchical Models Introduction
Model nested data structures — players within teams, shots within games — the most powerful Bayesian pattern you'll learn.
Common Approach Omits Information
Build forms that validate input and guide users to correct answers.
Multi-level Structure Propagates Uncertainty
Multi-level (hierarchical) models — the pattern for nested data structures in statistics.
Motivating Example
See why this technique matters — the real-world problems it solves that simpler methods don't.
§06
6. Sufficient Statistics
Sufficient Statistics Introduction
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Understanding Sufficient Statistics in Sports Analytics
Instrument apps to measure what users actually do — the input for product decisions.
Using Sufficient Statistics in Practice
Apply this technique to a real problem — theory becomes practice here.
§07
7. (More About) Correlation
(More About) Correlation Introudction
Measure relationships between variables and understand when correlation reflects real structure.
Trivariate Reduction
Model multiple outcomes jointly — richer than treating each outcome in isolation.
Marginal Plus Conditional
Work with marginal and conditional distributions together — the Bayesian workflow for complex models.
Copulas
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
§08
8. QR Decomposition
QR Reparameterization
Rewrite models to sample more efficiently — the technique that makes hard problems actually fit.
Understanding the Problem of Correlated Covariates
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Mathematics of QR Decomposition
Linear algebra under the hood — the math that makes statistical computation possible.
Practical Implementation in Stan and R
Implement the model in Stan — translate the math into code that actually fits the data.
§09
9. Autoregressive Processes
Autoregressive Processes Introduction
ROC curves and AUC — the classifier metrics that work across probability thresholds.
AR Processes with Equal Time between Measures
ROC curves and AUC — the classifier metrics that work across probability thresholds.
AR Processes with Irregular Times (Gaps) between Measures
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Coding the Model in Stan
Translate the mathematical structure directly into code — the step that makes models testable.
Fitting the Model with Data
Estimate your model's parameters and diagnose whether the fit is good enough to trust.
Multiple AR Processes with Interactions
ROC curves and AUC — the classifier metrics that work across probability thresholds.
§10
10. Survival Analysis
Survival Analysis Introduction
Survival analysis — model the time until an event happens, handling censoring correctly.
Time-to-Events | Proportional to Power Law
Survival analysis — model the time until an event happens, handling censoring correctly.
Simulate Time-to-Event Data
Work with this specific dataset and the modeling choices it motivates.
General Hazard and Survivor Functions
Write and use functions — the primary way to structure reusable, testable code.
Weibull Survival Model without Covariates
Survival analysis — model the time until an event happens, handling censoring correctly.
Weibull Model with Covariates
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Modifying the Priors for Baseball
Specify prior distributions that encode what you believe before seeing the data.
Fitting Model to Baseball Data
Apply the technique to real baseball data — historical, granular, and perfect for teaching concepts.
Posterior Inference
Extract insights from the posterior — the heart of every Bayesian analysis.
Interpreting the Coefficients
Turn raw model coefficients into insights stakeholders can act on.
Discretizing the Weibull Distribution
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Discrete Hazard Function
Write and use functions — the primary way to structure reusable, testable code.
Log-Hazard in Discrete Time
Model processes that advance in discrete time steps — useful for modeling state transitions.
Discrete Time Stan Model
Model processes that advance in discrete time steps — useful for modeling state transitions.
Estimate Parameters from Baseball Data
Fit model parameters from data — the core operation that every statistical model performs.
Posterior Predictive Checks
Validate the model by simulating data from it and comparing to observed reality. The essential sanity check.
§11
11. Differential Equations
Differential Equations Introduction
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Example: Usain Bolt's World Record Sprint
A worked example that demonstrates the technique on a concrete problem.
Example: Joint Model of Sprinter World Champions
Combine datasets with inner, left, right, and full joins — the SQL-style operations at the heart of analysis.
Using the Model
Apply this technique to a real problem — theory becomes practice here.
Refactoring the Model for Parallelizing the Likelihood
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
§12
12. Difference Equations
Difference Equations Introduction
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Modeling Performance: Bannister's Impulse-Response Model
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Cycling Power Data for Training
Apply the technique to cycling power and performance data — quantified athletic output.
Coding the Model in Stan
Translate the mathematical structure directly into code — the step that makes models testable.
Using the Model
Apply this technique to a real problem — theory becomes practice here.
§13
13. Splines
Splines Introduction
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Expected Goals Data
Work with this specific dataset and the modeling choices it motivates.
Simulating Data for Splines
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Model Overview
A high-level map of what this section covers and how the pieces fit together.
B-Spline Construction for a Single Variable
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Regression on B-Spline using Stan
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Counterfactual Data and use of Model
Reason about what would have happened under different conditions — the foundation of causal inference.
Speeding up the Spline
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Tensor Product of Spline Bases
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Two-Dimensional Splines
Smooth flexible curve fitting with splines — essential for non-linear relationships.
Implementing the Tensor Product in Stan and R
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Regression and Kroenecker Product
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Prediction for New Data
Use a trained model to generate predictions on new data.
Model Comparison
Compare competing models rigorously and pick the one that best explains the data.
Implications
The practical consequences of what you just learned — how it changes how you approach future problems.
§14
14. Gaussian Processes
Gaussian Processes Introduction
Flexible Bayesian function fitting with Gaussian processes — powerful when you need smooth priors over functions.
Likelihood
Specify how your data could have been generated under the model — the bridge from theory to evidence.
Gaussian Process Prior for Latent Function
Specify prior distributions that encode what you believe before seeing the data.
Understanding the Cholesky Decomposition in Gaussian Processes
Flexible Bayesian function fitting with Gaussian processes — powerful when you need smooth priors over functions.
Using the Cholesky Decomposition
Apply this technique to a real problem — theory becomes practice here.
Hyperparameters and Priors
Tune the knobs that control model behavior without leaking information from your test set.
Estimates at New or Counterfactual x
Reason about what would have happened under different conditions — the foundation of causal inference.
Stan Code for N-Dimensional Gaussian Processes
Flexible Bayesian function fitting with Gaussian processes — powerful when you need smooth priors over functions.
Testing the Code in One Dimension
Write tests for Shiny apps so you ship changes without breaking existing functionality.
Testing the Code in Two Dimensions
Write tests for Shiny apps so you ship changes without breaking existing functionality.
§15
15. Hilbert-Space Approximate GPs
Hilbert-Space Approximate GPs Introduction
Work with GPS tracking data — distances, speeds, acceleration profiles from wearable devices.
Math Refresher, Fourier Transforms
Build forms that validate input and guide users to correct answers.
HSGP Basis Functions Generally
Flexible function representations using basis functions — splines, polynomials, and beyond.
Frequencies in Basis Functions
Flexible function representations using basis functions — splines, polynomials, and beyond.
More Math, Spectral Densities
Density plots and estimation — visualize the full shape of a distribution, not just its summary.
Basis Functions
Flexible function representations using basis functions — splines, polynomials, and beyond.
Weighted Basis Function to Sum Model
Flexible function representations using basis functions — splines, polynomials, and beyond.
Likelihood for Observed Data
Work with this specific dataset and the modeling choices it motivates.
Priors for Parameters
Specify prior distributions that encode what you believe before seeing the data.
Full Model Specification
Declare models in a unified grammar that lets you swap algorithms without rewriting your pipeline.
Visual Walkthrough in r
The functional programming tools that replace most for-loops in modern R — faster and clearer.
Implementation of N-Dimensional HSGP in Stan
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
§16
16. Physics-Constrained Models
Physics-Constrained Models Introduction
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Sail GP Racing
Apply the modeling technique to competitive racing — a domain with rich sensor data and clear outcomes.
Load and Explore the Data
Work with this specific dataset and the modeling choices it motivates.
Physics of Sailing
Apply the modeling technique to competitive racing — a domain with rich sensor data and clear outcomes.
Coding the Physics in Stan
Translate the mathematical structure directly into code — the step that makes models testable.
Checking and Reviewing the Posterior
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Golf Putting
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Base Running
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Umpire Called Strikes
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
§17
17. Common Issues
Common Issues Introduction
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Outliers and Robustness
Identify and handle outliers — when they're errors, when they're signal, and when to leave them alone.
Missing Data Imputation
Detect, diagnose, and handle missing data without biasing your results.
Hit Tracking System Data in Baseball
Instrument apps to measure what users actually do — the input for product decisions.
Constraints in the Measurement Systems
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Mathematical Model
Build this model — the applied practice of probabilistic modeling.
Implementation in Stan
Implement the model in Stan — translate the math into code that actually fits the data.
Estimating Missing Values with the Model
Detect, diagnose, and handle missing data without biasing your results.
Censoring and Truncation
Handle censored and truncated data properly — critical for survival analysis and reliability.
Parameter Space Transformations
Build forms that validate input and guide users to correct answers.
§18
18. Computational Performance
Computational Performance Introduction
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Coding Optimizations
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Within Chain Parallelization
Parallelize computations to run faster on multi-core hardware.
GPU
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
Memory
Work through this lesson to build fluency with advanced Bayesian modeling, applied to complex hierarchical sports problems.
§19
19. Next Steps
Next Steps
Your roadmap for going further — the resources, packages, and patterns worth exploring next.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
03
Python MachineZ
Production ML in Python. scikit-learn, XGBoost, feature engineering, and the workflows that ship models to teams.
Taught by
Patrick McFarlane
Patrick McFarlane
§01
Introduction to Advanced Machine Learning in Sports Analytics
Machine Learning in Sports
Applied sports analytics context — the concrete domain where the techniques come to life.
Overview of Sports Data
A high-level map of what this section covers and how the pieces fit together.
Overview of Machine Learning
A high-level map of what this section covers and how the pieces fit together.
Syllabus Review
A structured preview of the course roadmap, graded assessments, and how sections connect.
QUIZ: Intro to Python MachineZ
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§02
Linear Regression and The Machine Learning Pipeline
Linear Regression and Least Squares
The foundation of supervised learning — fit, interpret, and understand when linear relationships hold.
Predicting PGA Tour Drive Distance
Use a trained model to generate predictions on new data.
Dataset Splits
Work through this lesson to build fluency with machine learning, applied to sports prediction and modeling.
Normalizing and Regularization
Control overfitting and handle high-dimensional problems with penalty terms that keep models honest.
Model Construction and Performance
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
QUIZ: Linear Regression & ML Pipeline
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
Exploratory Data Analysis and Feature Engineering
Motivating Data Exploration
See why this technique matters — the real-world problems it solves that simpler methods don't.
Data Visualization
Build publication-quality charts with the grammar of graphics — the strongest viz ecosystem in any language.
Feature Engineering
Decide which variables matter and how to transform them — often the difference between a good and great model.
QUIZ: EDA & Feature Engineering
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
Hyperparameter Tuning and Model Selection
Introduction to Hyperparameters
Tune the knobs that control model behavior without leaking information from your test set.
Cross Validation
Measure how well your model will generalize — the only honest way to compare algorithms.
Hyperparameter Tuning
Tune the knobs that control model behavior without leaking information from your test set.
Model Selection
Choose the right LLM for your task — performance, cost, and latency tradeoffs that actually matter.
QUIZ: HP Tuning and Model Selection
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
Model Performance Evaluation
Performance Metrics
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Choosing Evaluation Criteria
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Exploring Model Behavior
Explore the data systematically — the step that grounds every subsequent modeling choice.
QUIZ: Model Performance Evaluation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
Automating Machine Learning Pipelines
Motivating Machine Learning Pipelines
See why this technique matters — the real-world problems it solves that simpler methods don't.
Scikit-Learn Pipelines
Apply the dominant Python ML library to real problems — fit, predict, evaluate, and tune.
Pycaret
Work through this lesson to build fluency with machine learning, applied to sports prediction and modeling.
QUIZ: Automating ML Pipelines
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
Generalized Linear Models
Motivating Generalized Linear Models
See why this technique matters — the real-world problems it solves that simpler methods don't.
Outcome Distributions
Work through this lesson to build fluency with machine learning, applied to sports prediction and modeling.
Link Functions
Write and use functions — the primary way to structure reusable, testable code.
NFL Field Goal Probability
Probability as a language for uncertainty — the foundation for every statistical method.
QUIZ: Generalized Linear Models
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§08
Generalized Additive Models
Motivating Generalized Additive Models
See why this technique matters — the real-world problems it solves that simpler methods don't.
Basis Functions and B-Splines
Flexible function representations using basis functions — splines, polynomials, and beyond.
NBA In-Game Win Probability
Probability as a language for uncertainty — the foundation for every statistical method.
QUIZ: Generalized Additive Models
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§09
Tree-based Methods
Motivating Tree-based Methods
See why this technique matters — the real-world problems it solves that simpler methods don't.
Decision Trees
Tree-based models — intuitive, non-parametric, and the foundation for random forests and boosting.
Random Forest
Ensemble models that handle non-linear relationships and interactions without manual feature engineering.
Gradient Boosting Machines
Gradient boosting — how sequential models combine to build the strongest tabular learners available today.
WNBA Shot Probability
Probability as a language for uncertainty — the foundation for every statistical method.
QUIZ: Tree-Based Methods
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§10
Neural Networks
Motivating Neural Networks
Build and tune neural networks for problems where deep non-linearity matters.
Neural Network Architecture
Build and tune neural networks for problems where deep non-linearity matters.
Backpropagation
The algorithm that makes deep learning possible — compute gradients and update weights across any network.
Hyperparameters and Best Practices
Tune the knobs that control model behavior without leaking information from your test set.
Revisiting NBA In-Game Win Probability
Probability as a language for uncertainty — the foundation for every statistical method.
QUIZ: Neural Networks
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§11
Introduction to Applied Bayesian Statistics
Bayes' Theorem and Sampling
Understand how samples are drawn from populations and why sampling choices shape every inference.
Introduction to PyMC
Bayesian modeling in Python with PyMC — probabilistic programming without leaving the Python ecosystem.
Revisiting WNBA Shot Probability
Probability as a language for uncertainty — the foundation for every statistical method.
Wrapping Up
Closing thoughts and a summary of the skills you now have.
QUIZ: Intro to Applied Bayesian Stats
Test your understanding with targeted questions. Reinforce the concepts before moving on.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
04
Outstanding User Interfaces with Shiny: Mobile StructureZ
Mobile-first Shiny apps with shinyMobile. Touch gestures, native feel, and production-grade mobile architecture.
Taught by
Veerle Eeftink-van Leemput
Veerle Eeftink-van Leemput
Curriculum reviewed by
Dr. David Granjon
author of Outstanding User Interfaces with Shiny
Dr. David Granjon
§01
1. Introduction to the Course
Making Outstanding Shiny Apps
Orient yourself to what separates outstanding Shiny apps from typical ones.
Introducing Outstanding User Interfaces by David Granjon
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Prerequisites and Course Structure
Get your environment set up and confirm the knowledge you'll need before diving in.
§02
2. Introduction to Mobile Development for Shiny
Native Apps and Their Challenges
Cross the line from web app to app-like UX with mobile-native UI frameworks.
Introduction to Progressive Web Apps (PWAs)
Make Shiny apps installable on phones — offline support, push notifications, app-like experience.
Quiz: Progressive Web Apps
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Simulating Devices
Test mobile layouts without the slow dev loop — simulate any device from your browser.
§03
3. Behind {shinyMobile}
Introduction to Framework7
Cross the line from web app to app-like UX with mobile-native UI frameworks.
Setting up {shinyMobile}
Build true mobile-first Shiny apps with touch-optimized components and layouts.
Quiz: Setting up {shinyMobile}
Test your understanding with targeted questions. Reinforce the concepts before moving on.
About HTML Dependencies
Manage HTML assets properly in Shiny — the mechanism behind every custom widget.
Quiz: HTML Dependencies
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Framework7 Layouts
Cross the line from web app to app-like UX with mobile-native UI frameworks.
App Initialization
Work inside a real app — applied practice, not abstract theory.
Quiz: App Initialization
Test your understanding with targeted questions. Reinforce the concepts before moving on.
App Configuration: Theming and Colors
Work inside a real app — applied practice, not abstract theory.
Passing App Configuration from R to JS
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Quiz: App Configuration from R to JS
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Other App and Component Configuration
Work inside a real app — applied practice, not abstract theory.
Modularizing JS Code
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Quiz: Modularizing JS code
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Custom Message Handlers
Build custom components that go beyond what the defaults provide — the step toward polished work.
Quiz: Custom Message Handlers
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
4. Build a Mobile Sports App
The Data: Women's Basketball
Work through this lesson to build fluency with mobile-first Shiny apps, applied to touch-optimized apps for coaches and athletes.
Setting up the Layout
Work through this lesson to build fluency with mobile-first Shiny apps, applied to touch-optimized apps for coaches and athletes.
Building a Modular Codebase
Build it from scratch — the best way to internalize how the piece actually works.
Quiz: Building a Modular Codebase
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Ranking Page (Part 1)
Build this specific page of the app — a concrete exercise in assembling real UI.
Ranking Page (Part 2)
Build this specific page of the app — a concrete exercise in assembling real UI.
Quiz: Ranking Page
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Team Page (Part 1)
Build this specific page of the app — a concrete exercise in assembling real UI.
Team Page (Part 2)
Build this specific page of the app — a concrete exercise in assembling real UI.
Quiz: Team Page
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Player Page (Part 1)
Continue the walkthrough — advanced material that builds on the earlier parts.
Player Page (Part 2)
Continue the walkthrough — advanced material that builds on the earlier parts.
Login Functionality
Functional programming in R with purrr — map functions over lists, compose cleanly, skip the boilerplate.
Live Page Data
Work with this specific dataset and the modeling choices it motivates.
Home Page
Build this specific page of the app — a concrete exercise in assembling real UI.
Quiz: Home Page
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Creating a Messaging System
Work through this lesson to build fluency with mobile-first Shiny apps, applied to touch-optimized apps for coaches and athletes.
Quiz: Creating a Messaging System
Test your understanding with targeted questions. Reinforce the concepts before moving on.
User Settings and Preferences
Persist and apply user preferences across sessions.
Loading Experience
Work through this lesson to build fluency with mobile-first Shiny apps, applied to touch-optimized apps for coaches and athletes.
Quiz: Loading Experience
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
5. {shinyMobile} and PWA
{shinyMobile} and PWA concepts
Make Shiny apps installable on phones — offline support, push notifications, app-like experience.
Quiz: {shinyMobile} and PWA Concepts
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Manifest, Service Worker and Offline Page
Work through this lesson to build fluency with mobile-first Shiny apps, applied to touch-optimized apps for coaches and athletes.
Quiz: Manifest, Service Worker and Offline Page
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Google PWA Compatibility
Make Shiny apps installable on phones — offline support, push notifications, app-like experience.
Quiz: Google PWA Compatibility
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Shortcuts
The keyboard shortcuts that make serious R development dramatically faster.
Handling Installation
Install the app on users' devices — the PWA mechanics behind true mobile experiences.
To the App Store with Capacitor
Work inside a real app — applied practice, not abstract theory.
Quiz: Capacitor
Test your understanding with targeted questions. Reinforce the concepts before moving on.
To the App Store with PWABuilder
Make Shiny apps installable on phones — offline support, push notifications, app-like experience.
Quiz: PWABuilder
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
6. The Multilayout
The Framework7 Router
Cross the line from web app to app-like UX with mobile-native UI frameworks.
Quiz: The Framework7 Router
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Getting to Know {brochure}
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Multilayout in {shinyMobile}
Build true mobile-first Shiny apps with touch-optimized components and layouts.
Quiz: Multilayout
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
7. Recap and Next Steps
Summary
Consolidate the key concepts from the course and identify where to apply them in your own work.
What's Next?
Bridge to what comes after this course — the next skills, tools, or courses to accelerate your growth.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
● New Series · Launching Summer 2026
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
LESSON
NEW
Rapid App Prototyping with Claude CodeThe Agent Series · by Dr. Michael S. Czahor
The Agent Series
Build production sports tools using the modern agentic dev stack.
Claude Code, sub-agents, MCP, skills, automation, and live deployment — all taught through real sports analytics projects. 104 lessons shipping Summer 2026. Zero overlap with SynergiZing ML & LLMs in R; this series is about how you build, not what you model.
104+ lessons
10 sections
100% sports-centric
Posit Cloud VMs
Taught by
Dr. Michael S. Czahor
Dr. Michael S. Czahor
§01
Foundations of Agentic AI Development
The New Development Paradigm: AI as a Collaborator
Reframe your relationship with the editor — agentic development is a new profession, not just a tool.
Why Sports Analytics Is the Perfect Testbed for Agentic Work
Why sports data's richness, deadlines, and measurability make it the perfect domain to learn agentic development.
Terminal vs GUI: When Each Wins
When to reach for the terminal and when the GUI wins — the tradeoffs that shape every dev workflow.
Understanding IDEs and Their Role in AI Workflows
What an IDE actually does and why the choice matters more in the agentic era.
Getting Started with Visual Studio Code for Sports Work
VS Code set up for sports analytics — extensions, keybindings, and the sane defaults.
Setting Up Your Environment: R, Python, and Claude Code Together
Install R, Python, and Claude Code side-by-side so you can fluidly switch between analysis and agentic workflows.
Introduction to Claude Code: Your First Session
Your first Claude Code session on a real sports project — what to type, what to expect.
The CLAUDE.md File — Your Project Brain
The CLAUDE.md file is your project's memory — how to write one that keeps the agent aligned across long projects.
Anatomy of the .claude Directory
What lives in the .claude directory and why those configuration files matter for long-term project success.
Exploring Antigravity: When to Reach for It
When Antigravity helps, when it gets in the way, and how to integrate it into an existing workflow.
§02
Your First Agentic Build: A Scouting Report App
Project Brief: A Live MLB Scouting Dashboard
The capstone project of Section 2 — a live MLB scouting dashboard built from scratch with an agentic workflow.
Starting With a Clean CLAUDE.md for Sports Apps
The exact structure of a CLAUDE.md optimized for sports projects — data sources, quirks, conventions.
Approaches to App Design in Claude Code
Three approaches to designing an app in Claude Code and when each one produces the best results.
Iterating on UI with AI Collaboration
Iterate on UI with AI — how to give feedback that actually changes the output instead of re-prompting.
Pulling Sports Data: APIs, CSVs, and Databases
Pull sports data cleanly from APIs, CSVs, and databases — the agentic patterns that avoid flaky integrations.
Wiring Up the First Interactive Component
Wire your first interactive component end-to-end with agent-assisted development.
The Importance of Verification in AI-Built Code
Why verification matters more in AI-written code, and the habits that keep you safe from silent bugs.
Catching Hallucinations Before They Ship
Spot hallucinated function calls, made-up data sources, and imaginary APIs before they ship.
Your First Working Prototype — End to End
The full end-to-end build of your first prototype — code, data, UI, all agent-assisted.
Shipping the App Live
Actually deploy the prototype so a user can touch it — the step most AI tutorials skip.
§03
Context Management for Sports Projects
Context Management Explained
Context management is the single biggest factor in agentic dev productivity — how to think about it.
Structuring a Multi-Season Baseball Project
Structure a long project spanning multiple seasons so the agent always has the right context.
Managing Data Sources Across Sports
Multi-sport projects mean conflicting conventions — the patterns that keep the agent's mental model clean.
Using Scratchpads for Long Analyses
Use scratchpads for long analytical tangents without polluting the main project context.
Memory Strategies for Extended Work Sessions
Explicit memory strategies for extended work sessions — what to persist and what to discard.
When to Compact, When to Start Fresh
Compaction vs fresh start — the rule of thumb for long sessions that keeps costs and quality in balance.
Context Windows and Why They Matter for Production Work
How context windows actually work, and why the 200K limit shapes your entire production architecture.
Cost-Aware Development: Tracking Tokens in Real Projects
Track tokens in real projects and learn to estimate cost before starting an expensive agent run.
Pattern: The Daily Context Reset
The Daily Context Reset — a discipline that keeps long-running projects from drifting into chaos.
Pattern: The Long-Running Project
Run a project for weeks without the agent losing coherence — the patterns that actually scale.
§04
MCP (Model Context Protocol) for Sports Data
Introduction to Model Context Protocol
MCP is the plumbing that connects agents to your real tools — the quick orientation before you build.
Why MCP Changes How You Build Sports Tools
MCP unlocks tools that were previously impossible — the shift it represents for sports analytics.
Understanding MCP Servers
How MCP servers work under the hood — enough to debug them when they break.
Evaluating Token Usage in MCPs
MCP calls aren't free — how to evaluate and cap token usage per tool.
Connecting an MCP Server for Baseball Savant
Wire an MCP server to Baseball Savant so the agent can pull Statcast data on demand.
Connecting an MCP Server for NFL Play-by-Play
Connect an MCP server to NFL play-by-play data — instant granular queries from the agent.
MCP + Google Drive for Scouting Reports
MCP is the plumbing that connects agents to your real tools — the quick orientation before you build.
MCP + GitHub for Analytics Repos
MCP is the plumbing that connects agents to your real tools — the quick orientation before you build.
Building Your Own MCP Server: A Sports Stats Example
Build your own MCP server — a sports stats tool from the ground up.
Testing and Debugging MCP Integrations
The techniques that save hours when an MCP server silently stops working.
Deploying MCP Servers for Team Use
Ship your MCP server for team-wide use — auth, versioning, error handling.
Security Considerations for Sports Data MCPs
Security for sports-data MCPs — API keys, rate limits, and keeping proprietary data safe.
§05
Skills: Reusable Expertise for Sports Analysts
Introduction to Claude Skills
Skills turn one-off prompts into reusable expertise — the concept in one lesson.
The Anatomy of a Skill
The parts of a Skill and how each one shapes how the agent uses it.
SKILL.md Structure and Best Practices
Write a SKILL.md that consistently triggers when it should — and stays silent when it shouldn't.
Building Your First Skill: An xG Calculator
Build a Skill that wraps your xG calculation — the agent invokes it anytime xG comes up.
Building a Skill: Play-by-Play Parser
Build a Skill that parses raw play-by-play data into clean events the agent can reason about.
Building a Skill: Scouting Report Generator
Build a Skill that generates templated scouting reports from game-level input.
Building a Skill: Stan Model Validator
Build a Skill that validates Stan model diagnostics — R-hat, divergences, effective sample size.
Chaining Skills for Complex Workflows
Chain multiple Skills for complex workflows that no single prompt could handle reliably.
Versioning and Sharing Skills Across Your Team
How to version and share Skills across a team without chaos.
When Skills Are the Wrong Abstraction
Skills are powerful, but they're not always right — how to tell when a different abstraction fits better.
§06
Sub-Agents: Specialized Workers for Sports Tasks
Introduction to Sub-Agents
Sub-agents are specialized workers — the mental model before you start building them.
Turning a Skill into a Sub-Agent
When to promote a Skill to a Sub-agent — and how to do it cleanly.
Sub-Agent: Box Score Parser
A Sub-agent that parses box scores across leagues into a consistent schema.
Sub-Agent: Contract Analysis Specialist
A Sub-agent specialized in contract analysis — AAV, guarantees, cap hits, opt-outs.
Sub-Agent: Injury Report Classifier
A Sub-agent that classifies injury reports into severity, timeline, and return probability.
Sub-Agent: Draft Prospect Evaluator
A Sub-agent that evaluates draft prospects from multi-source scouting input.
Coordinating Multiple Sub-Agents
How multiple Sub-agents coordinate on one task — passing work between them without context loss.
When Sub-Agents Help (and When They Hurt)
The cases where Sub-agents help, and the cases where they just add complexity.
Debugging Sub-Agent Workflows
Debug failing Sub-agent workflows — the techniques that actually find the broken piece.
Token Efficiency Across Sub-Agent Hierarchies
Token efficiency across Sub-agent hierarchies — where costs hide and how to cap them.
§07
Agent Teams: Orchestrating Sports Workflows
Understanding Agent Teams
Agent teams are a step up from individual agents — the concept and when it matters.
Designing a Team for Draft Analysis
Design a team of agents for draft analysis — specialists that collaborate on evaluation.
Designing a Team for Live Game Monitoring
A team tuned for live game monitoring — low-latency agents that process events as they happen.
Designing a Team for Weekly Performance Reports
A weekly-reporting team that generates polished stakeholder summaries without manual work.
Enabling Agent Teams in Your Workflow
How to enable agent teams in your workflow and the tradeoffs that shape your setup.
Communication Patterns Between Agents
The communication patterns that keep multi-agent systems coherent.
Conflict Resolution When Agents Disagree
What to do when agents produce contradictory outputs — resolution patterns.
Cost Management at the Team Level
Cost management at the team level — per-agent quotas, sampling, and escalation.
When One Good Agent Beats a Team
Sometimes one strong agent beats a team of weaker specialists — how to tell the difference.
Case Study: Full Scouting Pipeline
A complete scouting pipeline case study — every agent, every handoff, every fallback.
§08
Automation and Git Worktrees for Sports Projects
The Power of Automation in Sports Analytics
Why automation pays off massively in sports work where games happen on strict schedules.
Utilizing Git Worktrees for Parallel Analyses
Use Git worktrees to run parallel analyses without branch-switching friction.
Automating Nightly Data Pulls
Automate nightly data pulls with agents that verify the data before trusting it.
Automating Weekly Report Generation
Automate weekly reports end-to-end — agents generate, verify, format, and send.
Scheduled Agent Runs: Cron, GitHub Actions, and Alternatives
Schedule agent runs with cron, GitHub Actions, or alternatives — tradeoffs of each.
Creating Automation Workflows That Actually Survive
Automation that actually survives for months — the failure modes to plan for.
Version Control Strategies for AI-Assisted Code
Version control strategies when a lot of your code is AI-assisted — commit hygiene for agents.
Code Review for Agent-Generated Pull Requests
Code review for agent-generated PRs — the checklist that catches issues humans miss.
Rollback Patterns When Agents Get It Wrong
Rollback patterns for when agents make it into main and you need to back them out quickly.
Audit Trails for Production Agent Decisions
Audit trails for agent decisions in production — what to log, what to ignore.
§09
Deployment and APIs for Live Sports Applications
Deployment Options for AI-Powered Sports Apps
The deployment options for agent-powered sports apps and when each makes sense.
Deploying APIs with Modal
Deploy a sports-stats API to Modal — serverless, fast to set up, scales on demand.
Serverless Deployment Patterns
Serverless patterns for agent-powered endpoints — cold starts, timeouts, and cost.
Building a Sports Stats API with Claude Code
Build a sports-stats API end-to-end in Claude Code — request parsing, auth, response shaping.
Authentication and Rate Limiting for Public Endpoints
Auth and rate limiting for public endpoints — the details that keep your API from getting crushed.
Monitoring Agent-Built Applications in Production
Monitor agent-built applications in production so you catch drift and failures early.
Scaling Agent Workflows for Peak Sports Seasons
Scale agent workflows for peak sports seasons — draft day, playoffs, trade deadline traffic.
Cost Optimization in Production
Cost optimization for agent apps in production — batching, caching, model selection.
Handling Live Game Data at Scale
Handle live game data at scale without blowing up latency or cost.
From Prototype to Production: A Full Walkthrough
The full walkthrough from agent-built prototype to a production system users rely on.
§10
Advanced Patterns and Case Studies
Gmail Integration for Recruiting Workflows
Integrate Gmail into a recruiting workflow — agents read, categorize, and draft replies.
Calendar Integration for Team Scheduling
Team scheduling via Calendar integration — agents handle travel, practices, and conflicts.
Slack Integration for Live Game Alerts
Slack integration for live game alerts — agents watch the data and notify the right channel.
Building a Custom Plugin for Claude Code
Build a custom plugin for Claude Code — extend the tool itself with your own capabilities.
Creating Specialized Evaluators for Sports Models
An applied lesson in the Agent Series — sports-centric agentic workflow taught by Dr. Michael S. Czahor.
Case Study: Full MLB Scouting System
Case study: a full MLB scouting system spanning pro, college, and international feeds.
Case Study: NBA Lineup Optimizer
Case study: an NBA lineup optimizer that accounts for matchups, fatigue, and rotation patterns.
Case Study: NFL Draft War Room Tool
Case study: an NFL draft war room tool that tracks board positions in real time.
Case Study: Soccer Transfer Market Analyzer
Case study: a soccer transfer market analyzer that models value, fit, and risk.
Case Study: Cycling Training Plan Generator
Case study: a cycling training plan generator that adapts to FTP, fatigue, and goals.
Common Agentic Development Anti-Patterns
The most common anti-patterns in agentic development — and the safer alternatives.
Debugging Strategies When Agents Go Sideways
Debugging strategies for when agents go sideways — triage framework and recovery paths.
This series is coming Summer 2026 and will be included with membership.
See Membership →
05
ProductioniZing Shiny Apps
Take Shiny apps from prototype to production. Modules, testing, deployment, and the patterns that survive real users.
Taught by
Veerle Eeftink-van Leemput
Veerle Eeftink-van Leemput
§01
Introduction to the Course
Welcome and Course Overview
Set expectations for the course, meet the instructor, and get oriented to the learning path ahead.
Prerequisites and Setting up your Environment
Get your environment set up and confirm the knowledge you'll need before diving in.
§02
What is Production?
Understanding Production in Software Development
What "production" actually means for Shiny — reliability, observability, and handling real users.
Considerations for a Production-Grade Shiny App
What "production" actually means for Shiny — reliability, observability, and handling real users.
Quiz: What is Production?
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
Building Your Foundation - A Simple Shiny App
The Shiny App Structure
The architecture behind a production Shiny app — UI, server, reactivity, and where things go.
Building the User Interface (UI)
Construct Shiny UIs that feel like real software, not a research demo.
The Server Logic
Write server-side code that's reactive, testable, and doesn't fall apart as the app grows.
Data, Functions and Constants: Basic Scoping Rules
Write and use functions — the primary way to structure reusable, testable code.
Quiz: Basic Scoping Rules
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Basic Reactivity
Master Shiny's reactivity model — the core mental model that makes advanced apps possible.
Quiz: Basic Reactivity
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Using Events
Apply this technique to a real problem — theory becomes practice here.
Quiz: Building Your Foundation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
Debugging Your Shiny App
Common Pitfalls and Troubleshooting
The mistakes most people make and how to avoid or diagnose them before they derail your project.
Debugging Strategies
Debug efficiently — the diagnostic techniques that separate frustrated devs from productive ones.
Quiz: Debugging Your Shiny App
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
Working with Data
Case Study - Olympics Data
Work with this specific dataset and the modeling choices it motivates.
Tabular Data
Work with this specific dataset and the modeling choices it motivates.
Interactive Graphics
Build charts that communicate insight instantly — visual design matters as much as the data.
Input Validation
Validate user input so bad data never reaches your backend or models.
Quiz: Input Validation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Handling File Uploads and Downloads
Handle file uploads in a way that scales — size limits, type checks, and progress indicators.
§06
Level-up your UI
Page Layout and Introduction to Bootstrap
Style your app with the bslib package — Bootstrap-powered theming that looks modern out of the box.
UI packages: bslib
Style your app with the bslib package — Bootstrap-powered theming that looks modern out of the box.
UI packages: bs4Dash
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
UI packages: shiny.semantic
UI packages that extend Shiny with richer component libraries and modern aesthetics.
Better Input Widgets
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Creating Dynamic UI Elements
Build rich UIs from Shiny's component library — the pieces you compose into apps.
Quiz: Creating Dynamic UI Elements
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
Advanced Reactivity
Understanding the Reactive Graph
Build charts that communicate insight instantly — visual design matters as much as the data.
Quiz: Understanding the Reactive Graph
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Observers, Outputs and Reactive Values
Handle side effects in Shiny with observers — the tools for actions that happen but don't return values.
Controlling and Stopping Reactivity
Master Shiny's reactivity model — the core mental model that makes advanced apps possible.
§08
Shiny Modules
Introduction to Shiny Modules
Structure large Shiny apps with modules — the pattern that keeps code maintainable as apps grow.
Building Your First Module
Build it from scratch — the best way to internalize how the piece actually works.
Quiz: Building Your First Module
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Reusing Modules Throughout Your App
Structure large Shiny apps with modules — the pattern that keeps code maintainable as apps grow.
Communication Between Modules (Sharing Data and State)
Structure large Shiny apps with modules — the pattern that keeps code maintainable as apps grow.
§09
Crafting a Seamless User Experience (UX)
Understanding UX and UI
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Designing and Prototyping
Design and prototype before writing code — the step that prevents massive rewrites later.
UX Design Principles
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Quiz: UX Design Principles
Test your understanding with targeted questions. Reinforce the concepts before moving on.
UI Design Patterns
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Quiz: UI Design Patterns
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§10
HTML and CSS for Shiny Apps
Basics of HTML
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Basics of (S)CSS
Style Shiny apps so they look like polished products, not research prototypes.
Using CSS: Finding Elements and Classes
Apply this technique to a real problem — theory becomes practice here.
Using CSS: Styling of UI Components
Build rich UIs from Shiny's component library — the pieces you compose into apps.
Using CSS: Transitions and Animations
Apply this technique to a real problem — theory becomes practice here.
Using CSS: Responsiveness
Apply this technique to a real problem — theory becomes practice here.
CSS Best Practices
Style Shiny apps so they look like polished products, not research prototypes.
Quiz: HTML and CSS for Shiny Apps
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§11
JavaScript
Why do we need JavaScript?
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Add Javascript to your Shiny App with shinyjs
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Write Custom JavaScript Code
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Quiz: Write Custom JavaScript Code
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§12
Best Practices in Shiny App Development
Structuring App Code
Work inside a real app — applied practice, not abstract theory.
Documentation
Write documentation that helps your future self and collaborators.
Dependency Management
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Version Control and Collaboration
Track changes and collaborate on code — the professional standard for all production work.
Quiz: Best Practices in Shiny App Development
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§13
Taking a Framework approach: Golem
Introduction to Golem
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Transforming your Shiny App to a Golem Shiny App
Work inside a real app — applied practice, not abstract theory.
Pros and Cons of Using Golem
The honest tradeoffs — every tool has them, and knowing them prevents painful surprises later.
Quiz: Golem
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§14
Streamlining with Rhino: Another Framework
Introduction to Rhino
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Transforming your Shiny app to Rhino Shiny App
Work inside a real app — applied practice, not abstract theory.
Pros and cons using Rhino
The honest tradeoffs — every tool has them, and knowing them prevents painful surprises later.
Quiz: Streamlining with Rhino
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§15
Advanced Techniques to Optimize Your Shiny App
Performance, Profiling and Benchmarking
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Caching
Cache expensive computations so users don't wait for work that's already been done.
A Gentle Introduction to Async Programming
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Tips For Efficient Code
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
§16
Security and Monitoring
Managing Secrets
Handle API keys and credentials securely — never commit them, never hardcode them.
Quiz: Managing Secrets
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Connecting with Databases
Connect Shiny apps to production databases — the pattern for apps that handle real data.
Data Security
Protect user data and credentials in production — the requirement every serious app must meet.
Quiz: Data Security
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Monitoring and Auditing
Instrument your production app so you know when something breaks and why.
§17
Testing
Introduction to Testing
Write tests for Shiny apps so you ship changes without breaking existing functionality.
Quiz: Introduction to Testing
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Unit Testing with testthat
Write tests for Shiny apps so you ship changes without breaking existing functionality.
End-to-end testing with shinytest2
Write tests for Shiny apps so you ship changes without breaking existing functionality.
§18
Deployment
Deployment Readiness
Move from RStudio to production — the deployment patterns that actually work at scale.
Quiz: Deployment Readiness
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Choosing the Right Deployment Option
Move from RStudio to production — the deployment patterns that actually work at scale.
Quiz: Choosing the Right Deployment Option
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Shinyapps.io
Deploy Shiny apps to hosting platforms teams actually use in production.
Shiny Server
Deploy Shiny apps to hosting platforms teams actually use in production.
Posit Connect
Deploy Shiny apps to hosting platforms teams actually use in production.
ShinyProxy
Deploy Shiny apps with ShinyProxy — containerized, authenticated, production-ready.
Heroku
Work through this lesson to build fluency with production Shiny applications, applied to real-world data products.
Serverless deployment with Shinylive
Move from RStudio to production — the deployment patterns that actually work at scale.
Continuous Integration and Continuous Deployment (CI/CD)
Move from RStudio to production — the deployment patterns that actually work at scale.
Quiz: CICD
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Automatic Deployment of the Olympic Games App to shinyapps.io
Move from RStudio to production — the deployment patterns that actually work at scale.
§19
Scaling Your App for Your Audience
Understanding Scalability Metrics: Sessions, Connections & Load Factor
Measure how your app handles concurrent users — the numbers that matter in production.
Quiz: Scalability Metrics
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Strategies for Scaling and Load Balancing
Handle real traffic with the right concurrency model — process workers, async, and more.
Quiz: Strategies for Scaling and Load Balancing
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§20
Course Recap and Next Steps
Course Summary
Consolidate the key concepts from the course and identify where to apply them in your own work.
What's Next? A Peek into Outstanding InterfaceZ in Shiny
Bridge to what comes after this course — the next skills, tools, or courses to accelerate your growth.
§21
Bonus Material
A Practical Example with renv: updating bslib
Style your app with the bslib package — Bootstrap-powered theming that looks modern out of the box.
Using Auth0
Secure your app with real user authentication — the production detail most tutorials skip.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
06
Outstanding User Interfaces with Shiny: CustomiZing Widgets
Custom HTML widgets, htmlwidgets, JavaScript integration — Shiny interfaces that feel like native web applications.
Taught by
Veerle Eeftink-van Leemput
Veerle Eeftink-van Leemput
Curriculum reviewed by
Dr. David Granjon
author of Outstanding User Interfaces with Shiny
Dr. David Granjon
§01
Introduction to the Course
Making Outstanding Shiny Apps
Orient yourself to what separates outstanding Shiny apps from typical ones.
Introducing Outstanding User Interfaces by David Granjon
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
§02
Package Development
The Why and How
Understand why this approach matters and exactly how to apply it in your own work.
Package Development Workflow
Turn your Shiny code into a proper R package — the step that separates hobbyists from professionals.
Quiz: Package Development
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
Web Application Concepts
The Client-Server Model and HTTP Requests
Understand how Shiny's client and server pieces communicate — essential for advanced UI work.
Quiz: The Client-Server Model
Test your understanding with targeted questions. Reinforce the concepts before moving on.
WebSockets
The protocol behind Shiny's real-time UI updates — understand it and you'll debug reactivity like a pro.
The Shiny App Lifecycle
Work inside a real app — applied practice, not abstract theory.
Quiz: The Shiny App Lifecycle
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
Discover Shiny Dependencies
Understanding Dependencies in Shiny Applications
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
Bootstrap
Style your app with the bslib package — Bootstrap-powered theming that looks modern out of the box.
jQuery
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
§05
Handle HTML dependencies with {htmltools}
Declaring and Attaching Dependencies
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
Resolving Dependency Conflicts
Resolve conflicting package versions cleanly — the silent killer of production deployments.
Quiz: Handle HTML Dependencies
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
JavaScript for Shiny
Introduction to JavaScript
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
jQuery Syntax
The language's syntax and how variables work — the starting point for every programming language.
Adding JavaScript to Shiny
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Quiz: Adding JavaScript to Shiny
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Using External JavaScript Libraries
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
§07
Communicate between R and JS
Data Exchange between R and JS
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Sending Messages through the WebSocket
The protocol behind Shiny's real-time UI updates — understand it and you'll debug reactivity like a pro.
Receiving Messages through the WebSocket
The protocol behind Shiny's real-time UI updates — understand it and you'll debug reactivity like a pro.
Quiz: Communication Between R and JS
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§08
Managing JS and CSS
The Need For Asset Management
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
Organize Your JS and CSS Files
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Bundling Your JS and CSS Code
Extend Shiny with JavaScript — the skill that separates functional apps from polished products.
Bundling Components
Bundle JavaScript and CSS assets for production Shiny apps.
Quiz: Bundling Components
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Working with npm
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
§09
Shiny Inputs
Input Bindings
Handle user input cleanly — validation, sanitization, and ensuring data flows work as expected.
Quiz: Input Bindings
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Updating Inputs
Work through this lesson to build fluency with custom Shiny UI components, applied to polished interactive dashboards.
Building A Complex Custom Input (Part 1)
Build inputs Shiny doesn't ship with — the pattern behind truly custom UIs.
Building A Complex Custom Input (Part 2)
Build inputs Shiny doesn't ship with — the pattern behind truly custom UIs.
Building A Complex Custom Input (Part 3)
Build inputs Shiny doesn't ship with — the pattern behind truly custom UIs.
Summarizing the Input Lifecycle
Understand how user inputs flow through Shiny — essential for debugging reactive chains.
Quiz: Summarizing the Shiny App Lifecycle
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§10
Shiny Outputs
Output Functions
Write and use functions — the primary way to structure reusable, testable code.
Output Bindings
Render computed outputs back to the UI — connect your server logic to what users see.
Quiz: Shiny Outputs
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§11
Mastering Shiny Events
Exploring Event Types
Explore the data systematically — the step that grounds every subsequent modeling choice.
Custom Loaders
Build custom components that go beyond what the defaults provide — the step toward polished work.
Quiz: Mastering Shiny Events
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§12
Design Widgets
Introduction to {htmlwidgets}
Build or extend custom HTML widgets that bring JavaScript libraries into Shiny seamlessly.
Building {brackets} with {htmlwidgets}
Build or extend custom HTML widgets that bring JavaScript libraries into Shiny seamlessly.
Advanced techniques in {htmlwidgets} (Part 1)
Build or extend custom HTML widgets that bring JavaScript libraries into Shiny seamlessly.
Advanced Techniques in {htmlwidgets} (Part 2)
Build or extend custom HTML widgets that bring JavaScript libraries into Shiny seamlessly.
Quiz: Advanced Techniques in {Htmlwidgets}
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Updating Your Widgets
Update widget state from the server — how components stay in sync with user actions.
§13
Course Recap and Next Steps
Course Summary
Consolidate the key concepts from the course and identify where to apply them in your own work.
What's Next?
Bridge to what comes after this course — the next skills, tools, or courses to accelerate your growth.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
07
BreeZing through the Tidyverse
The complete tidyverse walkthrough. dplyr, ggplot2, purrr, and the grammar of data science in R.
Taught by
Dr. Paul SabinEvan Callaghan
Dr. Paul SabinEvan Callaghan
§01
1. Course Introduction
Welcome and Course Overview
Set expectations for the course, meet the instructor, and get oriented to the learning path ahead.
Packages and Options
Configure R's packages and options for a smooth development experience.
Introducing Sports Datasets
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Basic R Code
The R fundamentals you'll use every day — syntax, data types, and the core verbs.
§02
Coding and Quizzes:
Coding Review: Course Introduction
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Course Introduction
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
2. Data Exploration
Atomic Data Types
Understand the primitive types R works with and why that affects every downstream operation.
Data Structures
Work with vectors, lists, matrices, and data frames — the containers your analysis runs on.
Reading in Data
Work with this specific dataset and the modeling choices it motivates.
Piping & Data Summaries
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
§04
Coding and Quizzes:
Coding Review: Atomic Data Types
Understand the primitive types R works with and why that affects every downstream operation.
Coding Review: Data Structures
Work with vectors, lists, matrices, and data frames — the containers your analysis runs on.
Coding Review: Reading in Data
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Piping & Data Summaries
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Data Exploration
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
3. Basics
Assigning Values in R
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Calling Functions in R
Write and use functions — the primary way to structure reusable, testable code.
Math Operations & Sequences
Math operations and sequences in R — the building blocks for numerical work.
§06
Coding and Quizzes:
Coding Review: Assigning Values in R
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Calling Functions in R
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Math Operations & Sequences
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Basics
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
4. Coding Style & Tidy Workflow
Spaces & Indentation
Code formatting conventions — the small choices that make codebases readable for years.
Pivot Longer
Reshape data between wide and long formats — the skill that unlocks most real-world analyses.
Pivot Wider
Reshape data between wide and long formats — the skill that unlocks most real-world analyses.
Nested Data
Work with this specific dataset and the modeling choices it motivates.
§08
Coding and Quizzes:
Coding Review: Spaces & Indentation
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Spaces & Indentation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Coding Review: Pivot Longer
Reshape data between wide and long formats — the skill that unlocks most real-world analyses.
Quiz: Pivot Longer
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Coding Review: Pivot Wider
Reshape data between wide and long formats — the skill that unlocks most real-world analyses.
Coding Review: Nested Data
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Nested Data
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§09
5. Loops & Vectorization
For Loops
Repeat operations efficiently — when to loop, when to vectorize, and how to know the difference.
While Loops
Repeat operations efficiently — when to loop, when to vectorize, and how to know the difference.
Vectorization
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
§10
Coding and Quizzes:
Coding Review: For Loops
Repeat operations efficiently — when to loop, when to vectorize, and how to know the difference.
Coding Review: While Loops
Repeat operations efficiently — when to loop, when to vectorize, and how to know the difference.
Coding Review: Vectorization
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Loops & Vectorization
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§11
6. Functions & Logic
If Statements
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Probability Functions
Write and use functions — the primary way to structure reusable, testable code.
Writing Functions
Write and use functions — the primary way to structure reusable, testable code.
§12
Coding and Quizzes:
Coding Review: If Statements
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Probability Functions
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Writing Functions
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Functions & Logic
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§13
7. Data Wrangling (Part 1)
Select
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Group By, Summarize, & Filter
The core dplyr verbs for splitting data into groups, summarizing each, and filtering to what matters.
Mutate
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
§14
Coding and Quizzes:
Coding Review: Select
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Group By, Summarize, and Filter
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Mutate
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Data Wrangling (Part 1)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§15
8. Data Wrangling (Part 2)
Tidyselect
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Across
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Missing Values
Detect, diagnose, and handle missing data without biasing your results.
§16
Coding and Quizzes:
Coding Review: Tidyselect
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Across
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Missing Values
Detect, diagnose, and handle missing data without biasing your results.
Quiz: Data Wrangling (Part 2)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§17
9. RStudio Shortcuts
RStudio Shortcuts
The keyboard shortcuts that make serious R development dramatically faster.
§18
Coding and Quizzes:
Coding Review: RStudio Shortcuts
Review the code patterns from the lesson and reinforce the implementation details.
§19
10. Data Visualization (Part 1)
Base R plots
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Intro to ggplot2
Build publication-quality charts with the grammar of graphics — the strongest viz ecosystem in any language.
Histograms
Histograms — the simplest honest way to visualize the distribution of a single variable.
Densities
Density plots and estimation — visualize the full shape of a distribution, not just its summary.
§20
Coding and Quizzes:
Coding Review: Base R Plots
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Intro to ggplot2
Build publication-quality charts with the grammar of graphics — the strongest viz ecosystem in any language.
Coding Review: Histograms
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Densities
Review the code patterns from the lesson and reinforce the implementation details.
§21
11. Data Visualization (Part 2)
Customizing Themes
Build custom components that go beyond what the defaults provide — the step toward polished work.
Colors, Palettes & Facets
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Team Logos & Colors
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
Calibration Plots
Calibration plots — visualize whether your predicted probabilities match observed frequencies.
§22
Coding and Quizzes:
Coding Review: Customizing Themes
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Colors, Palettes & Facets
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Team Logos & Colors
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Calibration Plots
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Data Visualization (Part 2)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§23
12. Factors
Intro to Factors
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
§24
Coding and Quizzes:
Coding Review: Factors
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Factors
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§25
13. Simulation (Part 1)
Setting up Simulation
Generate synthetic data from the model to test assumptions and validate downstream work.
Intro to data.tables
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
§26
Coding and Quizzes:
Coding Review: Setting up Simulation
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Intro to data.tables
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Simulation (Part 1)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§27
14. Simulation (Part 2)
Parallelization
Parallelize computations to run faster on multi-core hardware.
Rcpp Introduction
Call C++ from R for performance-critical code — the escape hatch when pure R is too slow.
§28
Coding and Quizzes:
Coding Review: Parallelization
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Rcpp Introduction
Review the code patterns from the lesson and reinforce the implementation details.
Quiz: Simulation (Part 2)
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§29
15. Strings
Intro to Strings
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
stringr
Work with text data using the string manipulation patterns every data professional needs.
Regular Expressions
Work with text data using the string manipulation patterns every data professional needs.
§30
Coding and Quizzes:
Coding Review: Intro to Strings
Review the code patterns from the lesson and reinforce the implementation details.
Coding Review: Stringr
Work with text data using the string manipulation patterns every data professional needs.
Coding Review: Regular Expressions
Work with text data using the string manipulation patterns every data professional needs.
Quiz: Strings
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§31
16. Dates/Times
Intro to Dates
Handle dates and times without pain — the package that makes R's weakest data type pleasant.
Intro to Lubridate
Handle dates and times without pain — the package that makes R's weakest data type pleasant.
§32
Coding and Quizzes:
Coding Review: Intro to Dates
Handle dates and times without pain — the package that makes R's weakest data type pleasant.
Coding Review: Lubridate
Handle dates and times without pain — the package that makes R's weakest data type pleasant.
§33
17. Joins
Joining Data in R
Combine datasets with inner, left, right, and full joins — the SQL-style operations at the heart of analysis.
§34
Coding and Quizzes:
Coding Review: Joining Data in R
Combine datasets with inner, left, right, and full joins — the SQL-style operations at the heart of analysis.
§35
18. Purrr
Intro to Purrr
Functional programming in R with purrr — map functions over lists, compose cleanly, skip the boilerplate.
Split Data Models
Work through this lesson to build fluency with tidyverse fundamentals, applied to real sports datasets including the NFL Draft.
§36
Coding and Quizzes:
Coding Review: Intro to Purrr
Functional programming in R with purrr — map functions over lists, compose cleanly, skip the boilerplate.
Coding Review: Split Data Models
Review the code patterns from the lesson and reinforce the implementation details.
§37
19. Communication
Intro to Quarto
Produce reproducible reports that combine code, results, and narrative in one document.
Quarto with Word and PowerPoint
Produce reproducible reports that combine code, results, and narrative in one document.
Additional Resources
Further reading and resources — where to go when you want to go deeper on your own.
§38
20. Case Study: NFL Draft Analysis
NFL Draft Curves and Data Loading
Load data cleanly from files, databases, and APIs — the unglamorous first mile of every project.
NFL Draft Pick Value Plots
Apply everything you've learned to real NFL Draft data — the capstone project that ties the course together.
NFL Draft Quarto Report
Produce reproducible reports that combine code, results, and narrative in one document.
NFL Draft & Trades Finalized Report
Apply everything you've learned to real NFL Draft data — the capstone project that ties the course together.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
08
FoundationZ of Data Science
Your Python on-ramp. NumPy, pandas, visualization, and the core toolkit every data scientist needs day one.
Taught by
Evan Callaghan
Evan Callaghan
§01
1. Course Introduction
Course Overview
A tour of what you'll build, the tools you'll use, and the real projects that tie the curriculum together.
What is Data Science?
A clear definition of the concept and why it matters for the work ahead.
Data Science in Sports
Work with the data — the ground truth that every model and analysis depends on.
Python for Data Science
Start writing Python for data science — the syntax and constructs you'll use daily.
§02
2. Python Basics
Introductory Python
Start writing Python for data science — the syntax and constructs you'll use daily.
Syntax and Variables
The language's syntax and how variables work — the starting point for every programming language.
Data Types and Structures
Work with the data — the ground truth that every model and analysis depends on.
Conditional Statements
Control program flow with if/else — the foundation of every non-trivial script.
Loops
Repeat operations efficiently — when to loop, when to vectorize, and how to know the difference.
Functions
Write and use functions — the primary way to structure reusable, testable code.
Libraries and Modules
Structure large Shiny apps with modules — the pattern that keeps code maintainable as apps grow.
Exception Handling
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Overview
A high-level map of what this section covers and how the pieces fit together.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
3. The NumPy Library
The NumPy Array
The core NumPy data structure — understand arrays to unlock the rest of scientific Python.
Generating Random Numbers
Build rating systems — the math behind Elo and similar competitive ranking approaches.
Array Operations
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Aggregation and Ufuncs
Summarize data at different levels of granularity — the core of every reporting workflow.
Filtering and Sorting
Filter data to the subset that's relevant — the most common reactive pattern in Shiny.
Array Manipulation
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Linear Algebra
Linear algebra under the hood — the math that makes statistical computation possible.
Overview
A high-level map of what this section covers and how the pieces fit together.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
4. The Pandas Library
Introduction to Pandas
Manipulate tabular data with pandas — the core tool for data work in Python.
Importing Data
Work with this specific dataset and the modeling choices it motivates.
Basic Data Exploration
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Filtering and Slicing DataFrames
Filter data to the subset that's relevant — the most common reactive pattern in Shiny.
Data Cleaning
Turn messy real-world data into the clean, model-ready format you actually need.
String Manipulation
Handle text data cleanly — parsing, splitting, matching, and transforming strings.
Aggregating and Summarizing Data
Work with this specific dataset and the modeling choices it motivates.
Data Transformations
Turn messy real-world data into the clean, model-ready format you actually need.
Combining and Merging
Combine datasets with inner, left, right, and full joins — the SQL-style operations at the heart of analysis.
Exporting Data
Work with this specific dataset and the modeling choices it motivates.
Overview
A high-level map of what this section covers and how the pieces fit together.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
5. Case Study: Data Wrangling
Case Study Overview
A high-level map of what this section covers and how the pieces fit together.
Data Wrangling
Turn messy real-world data into the clean, model-ready format you actually need.
Creating Grid Application
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
6. Data Visualization
Introduction to Data Visualization
Build publication-quality charts with the grammar of graphics — the strongest viz ecosystem in any language.
Anantomy of a Plot
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
The Matplotlib Library
Create charts and visualizations in Python — when to reach for each library.
Matplotlib Customization
Create charts and visualizations in Python — when to reach for each library.
The Seaborn Library
Create charts and visualizations in Python — when to reach for each library.
Seaborn Customization
Create charts and visualizations in Python — when to reach for each library.
Interactive Plots in Plotly
Create charts and visualizations in Python — when to reach for each library.
NHL Play-by-Play
Work with granular event-level sports data — the richest structure you'll encounter for modeling.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
7. Introduction to Modeling
Models Based on Similarity
Measure similarity between observations — the foundation of nearest-neighbor methods and clustering.
Regression vs. Classification
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Data Splitting
Work with the data — the ground truth that every model and analysis depends on.
The k-NN Algorithm
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
K-Means Clustering
Find structure in unlabeled data — segmentation, similar-player groupings, and pattern discovery.
Hierarchical Clustering
Model nested data structures — players within teams, shots within games — the most powerful Bayesian pattern you'll learn.
Exercises
Practice problems that move concepts from passive understanding to active skill.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§08
8. Case Study: Data Analysis
The EDA Process
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Defining the Problem
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Collecting our Data
Work with this specific dataset and the modeling choices it motivates.
Cleaning our Data
Work with this specific dataset and the modeling choices it motivates.
Aggregating our Data
Work with this specific dataset and the modeling choices it motivates.
Assessing Data Quality
Work through this lesson to build fluency with Python data science fundamentals, applied to sports analytics workflows.
Analyzing our Data
Apply analysis techniques to this dataset and extract the insights stakeholders actually need.
Modeling
Apply modeling techniques to this specific problem — the work of translating domain questions into models.
Clustering
Find structure in unlabeled data — segmentation, similar-player groupings, and pattern discovery.
Quiz
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§09
9. Looking Ahead
Course Recap
Review the full arc of the course and the skills you've added to your toolkit.
Machine Learning in Sports
Applied sports analytics context — the concrete domain where the techniques come to life.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
09
SynergiZing ML & LLMs in R
The modern R stack: tidymodels meets ellmer and ragnar. Tool-calling, RAG, agents, and production deployment.
Taught by
Dr. Nic Crane
Dr. Christoph Scheuch
Dr. Nic CraneDr. Christoph Scheuch
§01
1. Foundations of AI in R
ML vs LLMs
Understand where traditional ML ends and LLMs begin — and why modern work needs both.
AI Ecosystem in R
Tour the modern R AI stack — the packages and patterns production teams actually use.
§02
2. From Viz to Models
Play by Play Data
Work with this specific dataset and the modeling choices it motivates.
Exploration via Visualization
Build publication-quality charts with the grammar of graphics — the strongest viz ecosystem in any language.
Logistic Regression
Model binary outcomes with probabilistic interpretation — classification that actually tells you how confident it is.
§03
3. Model Evaluation and Testing
Training and Test Sets
Split data properly so your evaluation metrics actually predict real-world performance.
Model Specifications and Workflows
Declare models in a unified grammar that lets you swap algorithms without rewriting your pipeline.
Making Predictions
Generate predictions from trained models and understand what the outputs actually mean.
The Confusion Matrix and Basic Metrics
Go beyond accuracy to understand exactly how your classifier is succeeding and failing.
Advanced Evaluation Techniques
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Overfitting and Test Evaluation
Diagnose and fix the most common failure mode in machine learning — models that memorize instead of generalize.
§04
4. Resampling and Model Tuning
Introduction to Cross-Validation
Measure how well your model will generalize — the only honest way to compare algorithms.
Working with Resampling Results
Understand how samples are drawn from populations and why sampling choices shape every inference.
Comparing Models with Cross-Validation
Measure how well your model will generalize — the only honest way to compare algorithms.
Introduction to Hyperparameter Tuning
Tune the knobs that control model behavior without leaking information from your test set.
Selecting and Finalizing Models
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Tuning Multiple Parameters
Search the space of model configurations systematically and land on the best version for your data.
§05
5. Predicting Continuous Outcomes
From Classification to Regression
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Linear Regression in Tidymodels
The foundation of supervised learning — fit, interpret, and understand when linear relationships hold.
Random Forests for Regressions
Ensemble models that handle non-linear relationships and interactions without manual feature engineering.
Evaluating Regression Models
Measure whether your LLM actually works — eval datasets, scoring, and statistical comparison of models.
§06
6. Preprocessing with Recipes
Why Preprocessing Matters
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Essential Recipe Steps
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Preparing and Applying Recipes
Build reusable preprocessing pipelines that keep your feature engineering in sync with your modeling.
Recipes in Workflows
Bundle preprocessing and modeling into a single reproducible object.
§07
7. Regularized Regressions
The Problem of Overfitting
Diagnose and fix the most common failure mode in machine learning — models that memorize instead of generalize.
Ridge, Lasso, and Elastic Net
Control overfitting and handle high-dimensional problems with penalty terms that keep models honest.
Tuning Regularized Models
Search the space of model configurations systematically and land on the best version for your data.
Interpreting Regularized Coefficients
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
§08
8. Workflow Sets for Model Comparison
The Challenge of Fair Model Comparison
Compare competing models rigorously and pick the one that best explains the data.
Creating Workflow Sets
Compare multiple models in a fair, reproducible way with tidymodels' most powerful abstraction.
Tuning Workflow Sets
Compare multiple models in a fair, reproducible way with tidymodels' most powerful abstraction.
Selecting and Finalizing the Best Workflow
Bundle preprocessing and modeling into a single reproducible object.
§09
9. Neural Networks
Why Neural Networks?
Build and tune neural networks for problems where deep non-linearity matters.
Building Your First Neural Network
Build and tune neural networks for problems where deep non-linearity matters.
Tuning Neural Networks
Build and tune neural networks for problems where deep non-linearity matters.
Deeper Networks and Practical Considerations
Work with deeper neural networks — the practical considerations that matter as you stack layers.
§10
10. Deployment of Machine Learning Models
Why Deployment Matters?
Move from RStudio to production — the deployment patterns that actually work at scale.
Deploying Models with Vetiver
Move from RStudio to production — the deployment patterns that actually work at scale.
Exporting Models with Orbital
Export data and artifacts to formats your users actually work with.
Monitoring and Maintaining Deployed Models
Move from RStudio to production — the deployment patterns that actually work at scale.
§11
11. From ML to LLMs
A Gentle Introduction to LLMs
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Data Science Principles
Work with the data — the ground truth that every model and analysis depends on.
§12
12. Chat Basics
Introduction to LLMs
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Getting Started with Ellmer
Your first chat with an LLM from R — set up the client, send a prompt, parse a response.
Conversations, Roles, and Tokens
Understand how LLM conversations are structured and what tokens cost you in practice.
System Prompts
Shape LLM behavior with system prompts — the foundation of every production application.
§13
13. Prompt Engineering
Introduction to Prompt Engineering
Move from hoping-it-works to systematic prompt design that produces reliable output.
Adding Context and Instructions
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Handling Missing Data
Gracefully handle incomplete input and LLM output — production code has to handle the messy cases.
Fewshot Prompting
Teach the model by example — the prompting pattern that handles the most edge cases.
Structuring and Storing Prompts
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
§14
14. Structured Output
Introduction to Structured Output
Get LLMs to return clean data frames and structured JSON you can actually use in pipelines.
Extracting Data Frames
Turn unstructured text into structured data — invoices, reports, stats, whatever your pipeline needs.
Handling Missing Fields
Gracefully handle incomplete input and LLM output — production code has to handle the messy cases.
Batch Processing
Process thousands of prompts efficiently without blowing up your latency or API bill.
§15
15. Multimodal Input
Multimodal Input
Work with images, text, and tables together — extract structured data from visual sources.
Hallucination with Plot Interpretation
Understand why LLMs confabulate, when it matters most, and how to detect and reduce it in practice.
Extracting Data from Images
Turn unstructured text into structured data — invoices, reports, stats, whatever your pipeline needs.
§16
16. Model Selection
Model Selection
Choose the right LLM for your task — performance, cost, and latency tradeoffs that actually matter.
Working with Proprietary Models
Work with OpenAI, Anthropic, and Google models — when each is the right tool for the job.
Local Models Intro
Run open-source models on your own machine — privacy, cost control, no API dependency.
Running Local Models with Ollama
Run open-source models on your own machine — privacy, cost control, no API dependency.
Local Models in R
Run open-source models on your own machine — privacy, cost control, no API dependency.
§17
17. Tool Calling
Introduction to Tools
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Data Query Tools I
Register tools that let the LLM query your databases — the path to real data-driven AI assistants.
Data Query Tools II
Register tools that let the LLM query your databases — the path to real data-driven AI assistants.
Bio Retrieval Tools
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
User Control Safety
Give users control over what the system does on their behalf — the safety layer production systems need.
§18
18. NLP with Mall
Introduction to Mall
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Text Summarisation
Condense long documents to their essentials — a core LLM capability for reporting and research.
Extracting Information
Turn unstructured text into structured data — invoices, reports, stats, whatever your pipeline needs.
Text Classification
Classify text with LLMs at scale — often better than traditional classifiers, with less training data.
§19
19. Retrieval-Augmented Generation
Introducing RAG and Ragnar
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Document Processing
Turn raw PDFs and text into retrieval-ready chunks — the unglamorous but critical first step.
Retrieval via BM25
Classical keyword retrieval — still the strongest baseline for many document search tasks.
Retrieval via Embeddings
Convert text to vectors so you can find semantically similar content without exact keyword matches.
Hybrid Retrieval
Combine keyword and vector search for retrieval that beats either approach alone.
Registering Retrieval Tools
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Context Augmentation
Inject retrieved context into prompts in a way the LLM can actually reason about.
Multiple Documents
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
§20
20. Model Evaluation with Vitals
Introduction to Evals
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
Evaluation Data Sets
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Running First Evaluation
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Viewing Eval Results
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Comparing Models
Compare candidate models side-by-side so your choice rests on evidence, not intuition.
Statistical Comparison
Compare approaches side by side and understand when each one is the right tool.
RAG Evaluation Design
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Running RAG Evaluation
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
§21
21. LLM Apps
LLMs Data Exploration
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Text to SQL
Work through this lesson to build fluency with the R-native AI stack, applied to production LLM applications.
Querychat
Build natural-language-to-SQL interfaces — users ask questions, get answers, never write a query.
Querychat System Prompt
Shape LLM behavior with system prompts — the foundation of every production application.
Customizing Querychat
Build natural-language-to-SQL interfaces — users ask questions, get answers, never write a query.
Chatbots with Shinychat
Wrap your LLM logic in a production Shiny chat UI that users can actually interact with.
Tool Calling Part 1
Register R functions as tools the LLM can call — the pattern behind real AI agents.
Tool Calling Part 2
Register R functions as tools the LLM can call — the pattern behind real AI agents.
RAG in Chatbots
Build RAG pipelines that ground LLM responses in your actual documents — no more hallucinations.
Explaining Model Predictions
Interpret why your model made a specific prediction — critical for stakeholder trust and debugging.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
10
Biomechanics Tools and TechnologieZ
Force plates, motion capture, ForceDecks, IMU sensors — the data infrastructure of modern sports science.
Taught by
Dr. John Krzyszkowski
Dr. John Krzyszkowski
§01
1. Introduction to the Course
Welcome and Course Overview
Set expectations for the course, meet the instructor, and get oriented to the learning path ahead.
Resource Overview
Where to find the datasets, code repositories, and supplementary material that accompany the lessons.
Introduction to Python
Orient yourself to the topic, its vocabulary, and where it fits in the larger toolkit.
§02
2. Biomechanical Principles
Human Movement
The basics of how the body moves — joints, segments, and the vocabulary biomechanics uses.
Quiz: Human Movement
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Linear Kinematics
Describe motion without worrying about forces — positions, velocities, accelerations.
Angular Kinematics
Describe motion without worrying about forces — positions, velocities, accelerations.
Quiz: Kinematics
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Linear Kinetics
Measure the forces behind movement — the data that turns observation into intervention.
Angular Kinetics
Measure the forces behind movement — the data that turns observation into intervention.
Quiz: Kinetics
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§03
3. Motion Capture
Motion Capture Overview
Work with 3D motion capture data — the gold standard for biomechanics research.
Hardware Specifications
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Hardware Specifications
Test your understanding with targeted questions. Reinforce the concepts before moving on.
2D Motion Capture
Work with 3D motion capture data — the gold standard for biomechanics research.
3D Motion Capture (Part 1)
Work with 3D motion capture data — the gold standard for biomechanics research.
3D Motion Capture (Part 2)
Work with 3D motion capture data — the gold standard for biomechanics research.
Quiz: Motion Capture
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Acquisition of 3D Motion Capture (Part 1)
Work with 3D motion capture data — the gold standard for biomechanics research.
Acquisition of 3D Motion Capture (Part 2)
Work with 3D motion capture data — the gold standard for biomechanics research.
Calculating 3D Kinematic Data
Work with this specific dataset and the modeling choices it motivates.
Quiz: Calculating 3D Kinematic Data
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Motion Capture Applications
Work with 3D motion capture data — the gold standard for biomechanics research.
Quiz: Motion Capture Overview & Applications
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§04
4. Force Plates
Science Behind Force Plates
Measure the forces behind movement — the data that turns observation into intervention.
Hardware Specifications
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Hardware Specifications
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Setup and Calibration
Configure your biomechanics hardware and calibrate measurements so data is trustworthy downstream.
Quiz: Setup and Calibration
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Acquisition with Force Plates (Part 1)
Measure the forces behind movement — the data that turns observation into intervention.
Acquisition with Force Plates (Part 2)
Measure the forces behind movement — the data that turns observation into intervention.
Options for Force Plate Analysis
Measure the forces behind movement — the data that turns observation into intervention.
Quiz: Options for Force Plate Analysis
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Force Plate Applications
Measure the forces behind movement — the data that turns observation into intervention.
Quiz: Force Plates Overview & Applications
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§05
5. Other Sports Technologies
GPS Technology Overview
Work with GPS tracking data — distances, speeds, acceleration profiles from wearable devices.
EMG Technology Overview
Electromyography data — measure muscle activation and use it to analyze movement patterns and fatigue.
IMU Technology Overview
Inertial measurement units — accelerometer and gyroscope data from wearable sensors.
Sport Specific Technologies
Applied sports analytics context — the concrete domain where the techniques come to life.
Advanced Understanding through Integrations
Connect pieces together — the integration points that make systems work as a whole.
Quiz: Other Sports Technologies
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§06
6. Athlete Assessment Essentials
Why Assess Athletes
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Why Assess Athletes
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Movement Analysis
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Movement Analysis
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Performance Evaluation
Measure what matters for your problem — the right metrics depend on what you're actually trying to accomplish.
Quiz: Performance Evaluation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Athlete Monitoring
Build monitoring reports that track athlete workload, readiness, and performance over time.
Injury Risk and Rehabilitation
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Injury Risk and Rehabilitation
Test your understanding with targeted questions. Reinforce the concepts before moving on.
Advanced Modeling
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Quiz: Advanced Modeling
Test your understanding with targeted questions. Reinforce the concepts before moving on.
§07
7. Data Processing (Using Python)
Loading Biomechanical Data
Work with this specific dataset and the modeling choices it motivates.
Ingesting Supplemental Data
Work with this specific dataset and the modeling choices it motivates.
Data Preprocessing Techniques
ROC curves and AUC — the classifier metrics that work across probability thresholds.
Key Timepoint Identification
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
Plotting Time-Series Data
Work with this specific dataset and the modeling choices it motivates.
Key Performance Indicators
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Data Analysis
Work with the data — the ground truth that every model and analysis depends on.
§08
8. Report Building
Principles of Effective Report Building
Generate polished output documents — reports, PDFs, slide decks — from your analysis.
Individual Movement Report
Generate polished output documents — reports, PDFs, slide decks — from your analysis.
Longitudinal Performance-Based Report
Profile and optimize Shiny apps — where time actually goes and which tricks pay off.
Athlete Monitoring Report
Build monitoring reports that track athlete workload, readiness, and performance over time.
Injury Rehabilitation Report
Apply biomechanics to reduce injury risk and guide return-to-play decisions.
§09
9. Course Recap
Key Takeaways
The core lessons distilled — the ideas worth remembering long after the course ends.
Technologies in Sports Looking Ahead
Leave-one-out cross-validation — the Bayesian standard for comparing how well models generalize.
Identifying Your Niche / Send-Off
Work through this lesson to build fluency with biomechanics and motion analysis, applied to athlete performance and injury prevention.
This entire course is included with membership. Or enroll in just this course.
Course Details →
See Membership →
Nothing matches those filters.
Try widening your search or reset all filters.
Want all of that plus the live side?
AthlyticZ Membership launches July 2026 with the entire catalog above, 120+ masterclasses per year starting this summer, enterprise Posit Cloud VMs, and replays of every masterclass. Targeted workshops are available separately — and member discounts apply. One subscription. Everything you need to stay current.
Beyond Self-Paced
This doesn't even include
our live programming.
The catalog above is just the foundation. Starting this summer, AthlyticZ Membership unlocks a continuous flow of live, recurring content from instructors around the world — delivered fresh every month and archived for replay. Every masterclass ships with an enterprise Posit Cloud VM so you work in the same environment as the instructor.
120+
MASTERCLASSES / YEAR · LAUNCHING SUMMER 2026
Monthly live sessions with working practitioners
Roughly 10 masterclasses per month, starting this summer. Live Q&A, working sessions, code-alongs, career conversations. Attend live or watch the replay. Every session includes an enterprise Posit Cloud VM so you run the instructor's code in the same environment — no setup friction.
- Deep dives on current tools and techniques
- Live code reviews with experienced practitioners
- Career and interview prep sessions
- Q&A with instructors working at top teams and labs
- Enterprise Posit Cloud VMs included with every session
- Every session recorded and searchable in the member library
20+
GLOBAL INSTRUCTORS · 6 COUNTRIES · PURCHASED SEPARATELY
Targeted workshops on niche, high-value topics
Focused workshops led by instructors across pro sports, research, and industry — on topics you won't find on YouTube. Causal inference, Databricks, production AI prototyping, quant finance applications, interview prep — whatever you need to level up. Members receive discounts on all workshops.
- Niche topics: causal inference, APIs, D3.js, Quarto, SQL, interview prep
- Finance/quant, mobile apps, and domain-specific workshops
- Practitioners from MLB, NFL, NBA, academia, and industry
- Small-group settings with direct instructor access
- Purchased à la carte — members receive discounted pricing
- New workshops added continuously after July launch
Preview · Live Schedule Starts July
A taste of what's on the upcoming schedule.
Here's what live programming actually looks like. Instructors from six countries, shipping real sessions each month. Sample topics and formats shown below — real dates land when membership opens.
WS · 01
Workshop
4 hrs · 2 days
Refactoring Legacy R with Claude Code
Turn years of tangled R scripts into maintainable, tested code using agentic workflows. Hands-on, step-by-step.
Arkadi A.
R Production Specialist
WS · 02
Workshop Q&A
1 hr
Refactoring Legacy R — Live Q&A
Bring your legacy codebase and get direct, actionable guidance from a practitioner who ships R in production daily.
Arkadi A.
Included with Workshop
MC · 03
Masterclass
75 min
Choosing and Evaluating LLM Models
The decision framework for picking the right LLM for your task — cost, latency, capability, and eval strategy.
Dr. Nic Crane
15+ Years in R
MC · 04
Masterclass
75 min
Building a Live Scouting Dashboard
End-to-end build of a production-grade MLB scouting dashboard using modern R tooling, live with agentic co-development.
Dr. Michael S. Czahor
Founder, AthlyticZ
MC · 01
Masterclass
75 min
Natural Language Data Queries with querychat
Let stakeholders query your data in plain English. A live build using querychat, ellmer, and a production Shiny app.
Dr. Nic Crane
15+ Years in R
WS · 02
Workshop
3.5 hrs · 2 days
Large Language Models in R
The comprehensive R + LLM workshop. Prompting, tool-calling, RAG, evaluation, deployment. 7 hours of hands-on work.
Dr. Nic Crane
Flagship R+LLM Workshop
MC · 03
Masterclass
75 min
From Unstructured Text to Tidy Data
Turn PDFs, emails, and scraped text into analysis-ready data using LLM-assisted extraction patterns.
Dr. Nic Crane
15+ Years in R
WS · 04
Workshop
Coming Soon
Bayesian Modeling for Sports Decisions
From prior elicitation to posterior checks — Stan workflows that drive real decisions at pro sports organizations.
Dr. Scott Spencer
Columbia University
WS · 01
4-Part Workshop
4 hrs × 4 sessions
From Zero to Enterprise Deployment to SaaS
The full deployment spectrum. Local R scripts → Shiny Server → Posit Connect → production SaaS — the complete stack.
Stefan Linner
Enterprise R Deployment
MC · 02
Masterclass
75 min
Custom Shiny Widgets with htmlwidgets
Build Shiny components that feel like native web apps. JavaScript integration, D3, and the htmlwidgets framework.
Veerle Eeftink-van Leemput
Shiny Production Expert
MC · 03
Masterclass
75 min
Tidy Finance with R
The tidy approach to financial data, factor models, and portfolio analytics — by the co-creator of Tidy Finance.
Dr. Christoph Scheuch
Co-Creator, Tidy Finance
MC · 04
Masterclass
75 min
Production Stan at Scale
Scaling Stan models for high-throughput organizations — parallelization, reparameterization, and operational patterns.
Dr. Scott Spencer
Columbia University
Scroll to see more sessions
Stop piecing together YouTube tutorials.
One membership. Every course, every masterclass, every replay. Built by practitioners. Updated continuously. Designed for people who need to stay current.
Our Commitment to You
Nineteen focus areas.
And we're not slowing down.
Everything above is already live. We're adding hundreds of new project-based lessons over the next year — with a roadmap of thousands more beyond that. Taught through sports data, but the patterns transfer directly to finance, pharma, healthcare, marketing, and any industry that runs on data.
7 live now
2 just launched
10 in development
LIVE
Tidyverse
Data wrangling and analysis in R
LIVE
Shiny Apps
Production web applications in R
LIVE
Python Data Science
pandas, numpy, visualization, modeling
LIVE
Bayesian / Stan / PyMC
Probabilistic programming end-to-end
LIVE
Machine Learning
Classification, regression, deep learning
NEW
LLMs in R
Production R + LLM workflows
LIVE
Data Viz
ggplot2, plotly, custom charts
IN DEV
D3.js
Interactive web visualizations
LIVE
Sports Analytics
Every course taught with real sports data
IN DEV
APIs & Webscraping
Pull, parse, pipeline external data
IN DEV
Git Workflows
Version control for data teams
NEW
Agent Series / AI Prototyping
The Agent Series — Claude Code, sub-agents, MCP, automation
IN DEV
Quarto
Reproducible reports and dashboards
IN DEV
Causal Inference
Treatment effects, DAGs, synthetic controls
IN DEV
SQL
Query, join, window functions, optimization
IN DEV
Finance / Quant
Backtesting, factor models, portfolio math
IN DEV
Interview Prep
Case studies, SQL, modeling challenges
IN DEV
Mobile Apps
Touch-first Shiny and React Native patterns
IN DEV
Databricks
Spark, Delta Lake, production data platforms
The upskilling platform
for working data scientists.
Self-paced depth. Live masterclasses every month. Plus niche workshops with global practitioners (available separately). Built on the commercial tools companies actually use — not generic YouTube tutorials. Membership covers courses & masterclasses. Workshops available à la carte.
Membership launches July 2026All courses + masterclasses + Posit VMs
See Membership →