Where a good free book or guide already exists, this site links to it rather than rebuilding it. Ten of them, one line each on what it is good for and when to reach for it. All free, no courses to buy.
Last checked: October 2026The reference for pandas: loading, cleaning, groupby, merging, reshaping, time series, with the reasoning behind each. Use it when a Python video got you unstuck but you want to understand the tool properly; chapters 5 to 10 cover what the labs use.
Topic-by-topic explanations with runnable examples, more current than any book. Use it when you know what you want to do (merge, pivot, resample) and need the exact options; read the "10 minutes to pandas" page first if you are new.
Inference taught by simulation rather than formulas: sampling, bootstrap confidence intervals, hypothesis tests, then regression and classification. Use it when a p-value or a bootstrap interval in a lab feels like a ritual and you want to see where it comes from.
Short animated pages on probability, distributions, confidence intervals and regression that you play with rather than read. Use it for a ten-minute reset when a concept will not stick; it is a picture, not a course.
Economics done in code, from a Python-for-economists introduction up to dynamic programming and macro models, every result reproduced in a notebook. Use the introductory Python series if you want more practice after the labs, and the intermediate series if you are heading to graduate study or a research assistant post.
Every model and every evaluation tool in the library, each with the idea, the parameters and an example. Use it when a lab calls a model you have not met, or when you need to know what a parameter such as C or alpha is doing; the sections on model selection and cross-validation are the ones to read first.
The standard first textbook on supervised learning: regression, classification, resampling, regularization, trees, with Python labs for each chapter. Use it when you want the statistics behind a model rather than a recipe; the chapters on resampling and on linear model selection explain the bias-variance trade-off better than most lectures.
The econometrics of causal questions, written for economists: potential outcomes, matching, regression discontinuity, instrumental variables, difference-in-differences, synthetic control. Use it when a lab or a job asks you to estimate an effect rather than make a prediction, and you want the economist's version of the argument.
Starts from research design and causal diagrams and only then gets to the estimators, in plain language. Use it before the Mixtape if you have not taken econometrics, and use its chapters on fixed effects and event studies when you have a panel in front of you.
Four-hour courses on Python, pandas, data visualization, intro and intermediate machine learning, with exercises that check themselves. Use it when you want more reps with pandas or a first model between labs; skip the competitions until the course material is comfortable.
Every link was opened in October 2026 and pointed at the resource named. Links go stale; if one does not, let me know. Nothing here is paid or sponsored. For how these fit the jobs you are aiming at, see Careers with AI; for the skills each lab needs, see Labs and skills.