Editorial method
How MinutesRadar selects, summarises and publishes public council-minute records.
How records are made
- MinutesRadar collects public council minutes, extracts meeting text, identifies decisions, affected places, named projects and likely next steps, then generates crawlable pages from the structured record.
- Articles are drafted and checked with AI from public records. We do not imply that a reporter attended the meeting, conducted interviews or independently verified events beyond those records.
- Stories distinguish the recorded decision from analysis and questions for further reporting. A question is not an allegation, and an omission from minutes does not establish that information was withheld.
- New articles include quoted source passages checked against the extracted original text. Matching text is a traceability check, not a guarantee that the interpretation is correct. Stories with unresolved review flags are held back.
- Topic, area and project pages are generated from extracted evidence, not from manually invented keyword lists.
Quality controls
- The generator keeps weak area and project pages out of search indexes when there is not enough distinct local evidence yet.
- Broad issue matching is deliberately conservative so an address, legal act or generic construction word does not make one decision appear under unrelated topics.
- Pages include source links and dates so readers can check the record trail. Older articles may not have passage-level evidence or follow-up questions.
- Pen illustrations are AI-generated editorial artwork. They do not depict the actual people, buildings, sites or projects in a story.