Put the organizer-provided dataset files here (e.g. data/raw/).
⚠️ Everything in this directory except this README is gitignored — on purpose. Real crime records must never be pushed to a public repository, even "anonymized" ones, even temporarily. Treat the organizer's data-use terms as binding. Share files with teammates via a private channel, not git.
# 1. Profile every file — what's actually in there?
cd backend
python scripts/profile_dataset.py ..\data\raw\*
# 2. Read the generated report
# -> data/profile_report.md (rows, columns, nulls, date spans, geo candidates)
# 3. Then we write the ingestion adapter mapping real columns -> app/schema.py
# (PRD §9.1 — adapters, so schema surprises cost a day, not a week)No mock or synthetic data ships in this repo either — we don't guess the data format before seeing it. CI runs against an empty auto-created schema. Real data lives only in local/dev databases for analysis and the hosted demo, behind auth.