Is It A Bug Or Is It A Story?
summary
Shows concrete outcomes not airy promises: converts locked PDFs with PDFMiner and text‑mining to build a CC/BCC network that later proves to be a partial, cherry‑picked leak; scrapes FTP port manifests to trace US food‑aid shipments within a 10% error; ingests 2 GB CSV files of NHS prescriptions (70M rows yearly), discovers deanonymization at GP level and fixes by reaggregating to constituencies; packages work in VirtualBox and Vagrant for instant reproducibility and uses random‑forest decision trees for post‑hoc election analysis.
AI-written summary · about
chapters (19)
- 0:00Intro
- 1:22Nicolas background
- 2:25Nicolas story
- 9:32Government contracts
- 10:32Is it a bug
- 11:26The FIFA files
- 12:33Who thinks this is a story
- 13:39This is actually a bug
- 15:49This is a USbased story
- 16:32Building an interactive
- 17:36A legitimate story
- 18:44Large agribusinesses
- 19:02Naming companies
- 20:52Prescriptions data
- 24:25Its a bug
- 26:10Census data
- 27:02Predictions
- 27:23Linear regression
- 29:47Is it a story
related talks
06mUse Your Data Skills for Good
29mConquering PDFs- document understanding beyond plain text
35mBad data - anecdotes and examples from the real world
43mKeynote
27mExtracting tabular data from PDFs with Camelot & Excalibur
40mThe Paradise Papers: Behind the Scenes
52mUsing Open Source Tech to Unearth the Stories of the Panama Papers
51mFairytale vs. Narrative: Telling stories and making decisions with data