nyc-ride-market

L04 · Zones, congestion-zone list, airports, map geometry

Wave 1 · Phase 1 (draft), refined in 2 · Spec §3.3, §7, §9.2 (taxi_zones.geojson) · Decision D9 · Feeds L06, L07, L08, L10, L11, L14

Goal

A defensible list of taxi zones inside the Congestion Relief Zone, confirmed airport zones, and a simplified zone map under 1 MB.

Owns

• pipeline/zones.py, config/cbd_zones.csv, config/airports.csv, reports/cbd_zones.md, tests/test_zones.py
• Generated, not committed: data/ref/taxi_zones.simplified.geojson

Tasks

• Load taxi_zones.zip with DuckDB spatial; check its CRS and reproject to WGS84 for the GeoJSON
• Compare the lookup CSV with the shapefile: IDs 1–263 present, duplicates, name or borough mismatches; record findings in the report
• Draft cbd_zones.csv with columns LocationID, zone, borderline, note: Manhattan zones mostly south of 60th St. Write up every borderline call, with its reason, in reports/cbd_zones.md
• Helper cbd_zone_ids(zone_list="core" | "wide"): core excludes borderline zones, wide includes them (used by L06 and L07)
• Exploratory fee check on one 2025 HVFHV month, run on raw data in data/scratch/L04/: share of trips with cbd_congestion_fee > 0 by pickup and by drop-off zone. Use it to refine the borderline list; record the results
• Confirm 1,EWR, 132,JFK, 138,LGA against the lookup
• Simplify with mapshaper to under 1 MB, keeping LocationID, zone, borough

Tests

• Helper returns the expected IDs for both lists
• GeoJSON is under 1 MB and has 263 distinct LocationIDs (or the documented exceptions)
• Airports file matches the lookup

Done when

• Draft list and report merged for G1; refined after the exploratory fee check
• GeoJSON generated and under budget

Not this lane

• The formal SPEC §7.2 threshold check (L08)

---

Status log

Only this lane writes here. Newest first.

Date	State	Note
	not started

Blocked: none

Hand-off

• Built:
• Checked:
• Known limits:
• Follow-ups: