More fixes

pipeline/transform/poi_proximity.py

@@ -1,25 +1,29 @@
-"""Compute POI proximity counts per postcode from ArcGIS + filtered POIs."""
+"""Compute POI proximity counts and distances per postcode from ArcGIS + filtered POIs."""
 
 import argparse
 from pathlib import Path
 
 import polars as pl
 
-from pipeline.utils.poi_counts import count_pois_per_postcode
+from pipeline.utils.poi_counts import count_pois_per_postcode, min_distance_per_postcode
 
 
-# POI category groups for proximity counting.
+# POI category groups for proximity counting (2km radius).
 # Names must match the friendly names produced by transform_poi.py / naptan.py.
-POI_GROUPS = {
+POI_GROUPS_2KM = {
     "restaurants": ["Restaurant", "Fast Food"],
     "groceries": ["Greengrocer", "Supermarket", "Convenience Store"],
     "parks": ["Park"],
-    "public_transport": [
-        "Metro or Tram stop",
-        "Rail station",
-        "Bus stop",
-        "Bus station",
-    ],  # comes from naptan.py
+}
+
+# Train/tube stations counted at 1km radius
+TRAIN_TUBE_GROUP = {
+    "train_tube": ["Metro or Tram stop", "Rail station"],
+}
+
+# Groups for which to compute distance to nearest POI
+DISTANCE_GROUPS = {
+    "train_tube": ["Metro or Tram stop", "Rail station"],
 }
 
 
@@ -46,7 +50,21 @@ def main():
 
     pois = pl.read_parquet(args.pois)
 
-    result = count_pois_per_postcode(postcodes, pois, groups=POI_GROUPS, radius_km=2)
+    # Count amenity POIs within 2km
+    counts_2km = count_pois_per_postcode(
+        postcodes, pois, groups=POI_GROUPS_2KM, radius_km=2
+    )
+
+    # Count train/tube stations within 1km
+    counts_1km = count_pois_per_postcode(
+        postcodes, pois, groups=TRAIN_TUBE_GROUP, radius_km=1
+    )
+
+    # Distance to nearest train/tube station
+    distances = min_distance_per_postcode(postcodes, pois, groups=DISTANCE_GROUPS)
+
+    # Join all results on postcode
+    result = counts_2km.join(counts_1km, on="postcode").join(distances, on="postcode")
 
     result.write_parquet(args.output)
     size_mb = args.output.stat().st_size / (1024 * 1024)

pipeline/transform/postcode_boundaries/voronoi.py

@@ -16,8 +16,8 @@ def compute_voronoi_regions(
     if len(points) == 1:
         return {postcodes[0]: boundary}
 
-    # UPRN coordinates are int64 (BNG grid refs in whole meters).
-    # Convert to float64 so sub-meter jitter isn't truncated.
+    # UPRN coordinates are int64 (BNG grid refs in whole metres).
+    # Convert to float64 so sub-metre jitter isn't truncated.
     points = points.astype(np.float64)
 
     # Deduplicate points, keeping one per (location, postcode) pair.