Day 1: Lightning Talks
summary
A rapid sequence of concrete demos and claims: a tour of the NumFOCUS project directory and the difference between sponsored vs affiliated projects; a model‑explainability demo using the alibi library with anchors with 92% precision and counterfactuals on Income, Boston Housing, FashionMNIST and ImageNet; spatial crime mapping built with GeoPandas, Plotly and Mapbox; interactive dashboards via Voila, ipywidgets and bqplot; reproducible dev flow using pre‑commit hooks, black and isort; and an elections dataset package plus a ML entry that reduced error from 4% to 0.02%.
AI-written summary · about
chapters (12)
- 0:00NumFOCUS projects
- 5:49Model explainability with Alibi
- 11:10UK crime data visualization
- 16:18PyData Cambridge 2023
- 17:44Python and CircuitPython fun
- 25:25Git hooks for Python
- 30:07Jupyter dashboards with Voila
- 35:14SciPy 2023 in Bilbao
- 36:45PyCon DE & PyData Berlin 2023
- 39:54Python quirks
- 45:31Democracy hackathon results
- 51:09Bird classification with widgets
