// the find
justmarkham/DAT8
General Assembly's 2015 Data Science course in Washington, DC
Course materials from a 2015 General Assembly data science bootcamp taught in Washington DC. It's a 22-class curriculum covering pandas, scikit-learn, and the standard ML toolkit — KNN, linear/logistic regression, decision trees, naive Bayes, clustering, NLP, and ensembles. Aimed at beginners transitioning into data science, not working practitioners.
The notebook quality is genuinely good — each topic has worked examples on real datasets (Titanic, bikeshare, IMDb) rather than toy arrays. The resource lists per class are thorough and point to legitimately useful material like Hastie & Tibshirani and the Caltech learning course. The curriculum sequencing is sensible: EDA before modeling, model evaluation before moving to more complex algorithms. Having homework solutions checked in alongside assignments is useful for self-study.
This is a 2015 course and it shows — Python 2.7, Anaconda installs via conda, scikit-learn APIs that have since changed. Anyone running this today will hit deprecation warnings or outright breakage. There's no deep learning content at all, which was already becoming relevant in 2015 and is now table stakes. The course assumes instructor-led delivery; someone working through it solo loses the homework submission forms, Slack channel, and peer review that structured the pacing. At 1,600 stars and 1,000 forks, most of the interest is archival — the DataSchool blog is the better ongoing resource from the same instructor.