Stop Publishing Garbage Data, It's Embarrassing
6 hours ago
- #Data Validation
- #Data Quality
- #Government Data
- Two instances of poor-quality data were encountered, highlighting issues in UK government and organization datasets.
- The UK government's fuel finder CSV file contained errors such as incorrect latitude/longitude placements and unrealistic fuel price ratios.
- Despite reporting the errors on March 22, 2026, the data remained unchanged in the updated file published on March 29, 2026.
- A report from the RAC included a graph showing an implausible drop in Battery Electric Vehicles, likely due to a units mix-up.
- Bad data undermines trust and can lead to poor decisions, with concerns about future data pollution from unchecked LLM-generated data.
- A call for proper validation and proofreading in data handling to maintain quality and prevent a 'slop-apocalypse'.