Repository navigation
Collect pilot metadata and join it to telemetry for clustering #214
Description
Is your feature request related to a problem? Please describe.
The client wants pilot rating and flight hours to factor into clustering. Neither exists anywhere in the repo, and there is no place to put them.
Describe the solution you'd like
A small CSV keyed by pilot_id with certification_level, flight_hours and pilot_rating, plus a loader in inference/ that joins it to the telemetry at feature engineering time. Handle missing pilots without dropping their telemetry silently.
Additional context
Field definitions are in docs/telemetry_schema.md section 4b. Depends on #211 for the join key. The category set for certification_level (Student, Private, Commercial, ATP) still needs confirming, and so does whether pilot_rating is a licensing rating or an evaluative score, which changes its column type. Both questions are with the client.
Activity
jmacd867 commentedon Oct 3, 2026
certification_levelagainst an ordered category list and map it to an ordinal 0-3 (Student, Private, Commercial, ATP). The set is still to be confirmed with the client.pilot_ratingas a string and keep it out of clustering until the client says whether it is a licensing rating or an evaluative score. Keep a single clearly named place to change that.Phlabry commentedon Oct 5, 2026
@jmacd867 the join key is in review as #219. Columns are
pilot_idandflight_id, written bydata_logger.pyinto every row, so you can build the join against those names now.jmacd867 commentedon Oct 16, 2026
Answers from the client meeting (`docs/client_meeting_notes.md`):
The `certification_level` category set (Student, Private, Commercial, ATP) is still to be confirmed.