PyData London 2017, Lightning Talks - Saturday
summary
Collection of five rapid demos shows concrete claims and tools: Facebook Prophet produces probabilistic daily forecasts with changepoints and uncertainty intervals via a scikit‑learn style API; nbdime diffs and intelligently merges Jupyter notebooks (preserves valid JSON after conflicts); batch runs use runipy/Jupyter‑stream to parameterize notebooks; experiments with Google's Inception model reveal surprising dog vs bread misclassifications; a tweet bot scrapes race timing into SQLite and posts updates; a meta‑makefile plus UnionFS builds 464 packages reproducibly.
AI-written summary · about
chapters (10)
- 0:00Welcome to PyData London 2017
- 0:21Probabilistic time series with Prophet
- 4:35Improving Jupyter with nbdiff/nbmerge
- 8:54Automating workflows with Running-Pi
- 12:06Testing Google’s captcha recognition
- 16:47Live half-marathon Twitter bot
- 21:22Python 3 migration in Jupyter
- 26:40Automated notebook testing
- 27:44Managing code with meta makefile
- 35:05Closing remarks
