Democratising data journalism
summary
Bureau Local builds a collaborative newsroom that turns national open datasets into regional investigations using Python, publishing 169 local stories and 40+ open resources from a 700‑strong network. Examples: a model that forecasts snap election swings to ~1% accuracy, analysis of Facebook dark ads using the Who Targets Me Chrome plugin, centralised council budgets and council reserves datasets, and an automation that commits a Google sheet to GitHub after edits — data pipelines that run in 22 seconds to merge 540 CSVs.
AI-written summary · about
chapters (23)
- 0:00Bureau Local team intro
- 1:53Origins of Bureau of Investigative Journalism
- 3:18Local journalism needs data collaboration
- 4:43Data flaws in UK election system
- 6:27What data journalism means
- 8:15Data as lead not story
- 9:19Making tech skills accessible
- 11:09Barriers local reporters face
- 13:42Automating data processing
- 15:33Why collaboration fails
- 16:35Reporting recipes for journalists
- 17:45Open data vs FOI requests
- 19:50Bureau Local’s snap election response
- 21:58Predicting election outcomes locally
- 24:50Hosting data days across UK
- 26:58Election investigation results
- 27:40Immigration profiling patterns
- 29:46US vs German data journalism
- 31:10Impact of immigration investigation
- 32:12Domestic violence refuge crisis
- 35:05Centralizing council budgets
- 37:30Tracking Alma’s death cases
- 38:56Bureau Local growth and impact
