A data quality case study from English local elections on categorical normalisation, metric validation, and why raw labels should never define analytical groups. The post Churn Without Fragmentation: How a Party-Label Bug Reversed My Headline Finding appeared first on Towards Data Science.