Statistics
Statistics turns observations into claims you can test. Start with variable types, learn to read distributions, then summarize them without hiding their shape.
The track's full scope — sampling, estimation, hypothesis testing, regression, Bayesian thinking, experiment design, causal inference, time series — is outlined in the curriculum.
In review — 1 draft awaiting publication
Planned lessons (30)
Wave 1 — first vertical slice (production-ready) — 30 lessons
foundations
- Summary Statistics and Their Traps Means that lie, medians that hide, and the statistics that survive skew.
- Probability Rules Without Tears Sample spaces, conditioning, and Bayes' rule without the ritual.
- Random Variables Random variables, expectation, variance: the vocabulary of every model.
- The Distributions You'll Meet Bernoulli to normal: where each distribution comes from and what it describes.
inference
- Sampling Samples, populations, and the bridge between them.
- Sampling Distributions and the CLT Why the mean of almost anything settles down.
- Estimation: Bias and Variance What a good estimator owes you, and the trade it makes.
- Confidence Intervals, Honestly What a CI actually claims, and the confidence it doesn't.
- The Hypothesis Testing Framework Null hypotheses, p-values, and the logic of the test.
- t-Tests One- and two-sample t-tests, and the normality they quietly assume.
- Chi-Square Tests Categories under test: goodness of fit and independence.
- ANOVA Comparing several means honestly, and what to do after the F-test.
- Nonparametric Tests Rank-based tests for when the assumptions won't hold.
- Multiple Testing Why 20 tests at p<0.05 guarantee lies, and the corrections that fix it.
regression
- Correlation and Its Discontents What correlation measures, and the stories it can't support.
- Simple Linear Regression One predictor, one line: fitting, reading, and doubting the slope.
- Multiple Regression Coefficients as adjusted effects, and the moment they stop meaning that.
- Regression Diagnostics Residuals, leverage, influence: the regression's confession.
- Logistic Regression Modeling probabilities with a linear heart.
bayes
- Bayesian Thinking Priors, likelihoods, posteriors: belief updated by evidence.
- Bayesian Analysis in Practice A real Bayesian analysis in PyMC or Stan, start to posterior.
experiments
- A/B Testing: Design Randomization, units, and metrics: an experiment that actually answers.
- A/B Testing: Analysis Effects, intervals, and the traps between them.
- Statistical Power Designing experiments that can detect the effect you care about.
- Sequential Testing Peeking, early stopping, and the corrections that keep peeking honest.
causal
- The Causal Ladder Association, intervention, counterfactuals: the three rungs of causation.
- Confounding and DAGs Draw the causal graph, find the confounders, adjust for the right ones.
- Quasi-Experiments Difference-in-differences and regression discontinuity: causation without labs.
time-series
- Time Series Decomposition Trend, seasonality, remainder: taking a series apart.
- Forecasting Baselines Naive and seasonal baselines, and the humility they enforce.