A data science team's actual lunch-and-learn notes, open-sourced
IPython notebooks from Twitter Boulder's data science team, covering everything from pandas to neural networks in bite-sized weekly sessions.

What it does This repo collects the hands-on teaching materials from a working data science team’s weekly 45-minute learning sessions. Topics span Python basics, statistics, machine learning, NLP, network analysis, visualization, and engineering practices — all presented as notebooks, READMEs, or interactive walkthroughs.
The interesting bit These aren’t polished courses; they’re working notes from practitioners teaching each other. The breadth is unusual — you’ll find count-min sketches next to Flask basics, causal inference beside bash data structures. It’s a rare look at what a team actually decided was worth learning together.
Key highlights
- 50+ topics across Python, stats, ML, NLP, networks, viz, databases, and engineering
- Heavy emphasis on hands-on formats: IPython notebooks, knitr, live coding sessions
- Includes niche practical topics often skipped in formal curricula:
jq, Vertica, horizon charts, bandit algorithms - Originated at Gnip/Twitter Boulder, so social data and streaming contexts appear throughout
- Explicitly welcomes pull requests and reuse for other teams
Caveats
- Quality and depth vary by topic; these are team notes, not vetted courses
- Some links may be stale; the README doesn’t indicate last-updated dates
- “Every week*” comes with an asterisk and guilt trips, so consistency is unclear
Verdict Great for data scientists or Python developers building a self-study curriculum or team training program. Less useful if you need polished, sequential coursework — this is a grab bag, not a degree.