Better

pipeline/transform/poi_proximity.py

@@ -56,6 +56,22 @@ DYNAMIC_FILTER_ALL_GROUPS = {"Public Transport", "Leisure", "Health"}
 DYNAMIC_FILTER_COUNT_THRESHOLD_GROUPS = {"Groceries"}
 DYNAMIC_FILTER_EXCLUDED_CATEGORIES = {"Park"}
 
+# Combined "nearest of any rail mode" distance: the distance to the closest of a
+# Rail station, Tube station, DLR station, or Tram & Metro stop. Materialized as
+# its own real column so the server loads it like any other per-category
+# distance. min_distance_per_postcode already supports multi-category groups, so
+# this is a genuine nearest-of-the-union query (identical to the per-mode
+# minimum). Distance only: a combined "within Xkm" count is not meaningful, so
+# it is intentionally omitted from the count groups.
+COMBINED_STATION_GROUP_KEY = "poi_any_station"
+COMBINED_STATION_DISPLAY_NAME = "Any station"
+COMBINED_STATION_SOURCE_CATEGORIES = [
+    "Rail station",
+    "Tube station",
+    "DLR station",
+    "Tram & Metro stop",
+]
+
 
 def _poi_category_slug(category: str) -> str:
     ascii_text = (
@@ -227,10 +243,21 @@ def main():
     dynamic_counts_5km = count_pois_per_postcode(
         postcodes, pois, groups=poi_category_groups, radius_km=5
     )
+    # Distances additionally include the combined "Any station" group (nearest of
+    # any rail mode). It is distance-only, so it is added here but not to the
+    # 2km/5km count groups above.
+    distance_groups = {
+        **poi_category_groups,
+        COMBINED_STATION_GROUP_KEY: COMBINED_STATION_SOURCE_CATEGORIES,
+    }
+    distance_display_names = {
+        **poi_display_names,
+        COMBINED_STATION_GROUP_KEY: COMBINED_STATION_DISPLAY_NAME,
+    }
     dynamic_distances = min_distance_per_postcode(
-        postcodes, pois, groups=poi_category_groups
+        postcodes, pois, groups=distance_groups
     )
-    dynamic_renames = _dynamic_poi_metric_renames(poi_display_names)
+    dynamic_renames = _dynamic_poi_metric_renames(distance_display_names)
     dynamic_counts_2km = dynamic_counts_2km.rename(
         {k: v for k, v in dynamic_renames.items() if k in dynamic_counts_2km.columns}
     )