Data Sources¶
1. Boundary & Reference Data¶
- Neighborhood boundaries
- Source: OpenDataPhilly
- Filename: philly_neighborhoods.geojson
- Used for: aggregation, mapping, and neighborhood-level scoring
2. Environmental Data¶
-
NDVI (Vegetation Health)
- Source: Google Earth Engine (Landsat 8 SR)
- Processing: seasonal filtering, NDVI formula, zonal statistics
- Normalized output column:
ndvi_score
-
Tree Canopy (PPR 2015 Canopy Points)
- Source: Philadelphia Parks & Recreation
- Processing: point density → normalized 0–1
- Output column:
tree_score
-
Park Proximity (OSM Parks)
- Source: OpenStreetMap (
leisure=park,garden,playground) - Method: KDTree nearest-neighbor distance → inverted → score
- Output column:
park_proximity_score
- Source: OpenStreetMap (
-
Final environmental score column:
environmental_score = (ndvi_score + tree_score + park_proximity_score) / 3
3. Land Use & Amenity Data¶
- Healthcare locations (OSM)
-
Tags: hospital, clinic, pharmacy, dentist, emergency, nursing_home
- Method: KDTree nearest facility → score
- Output column:
health_proximity_score
-
Civic & Social Services (OSM)
- Tags: social_facility, post_office, library, police, fire_station, childcare
- Method: KDTree nearest service → score
- Output column:
service_proximity_score
-
Final land-use score column:
land_use_score = (health_proximity_score + service_proximity_score) / 2
4. Mobility & Public Realm Data¶
-
Sidewalk Network (OSM Footways)
- Tags: highway=footway, footway=sidewalk, crossing
- Method: NetworkX components → gap density → normalized
- Output column:
sidewalk_index
-
Cartways & Curbs (OpenDataPhilly)
- Files: Curbs.geojson + Curbs_No_Cartways.geojson
- Method: cartway density → inverted → normalized
- Output column:
cartway_density_index
-
Bike Network (OpenDataPhilly)
- File: Bike_Network.geojson
-
Method: lane density normalized
- Output column:
bike_index
- Output column:
-
Final mobility score column:
mobility_score = (bike_index + sidewalk_index + cartway_density_index)/3
5. Social Demographic Data (ACS 2022)¶
- Variables used:
- Disability rate (inverted)
- Elderly 65+ rate (inverted)
- Walk-to-work share
- Level: census tract, aggregated to neighborhoods
- Output column:
social_score
6. Composite Accessibility Score¶
- Weighted formula used in analysis:
\[
\text{Accessibility Score} =
0.4 \times \text{Mobility Score} +
0.3 \times \text{Land Use Score} +
0.2 \times \text{Environmental Score} +
0.1 \times \text{Social Score}
\]
- Output column:
accessibility_score
7. Quick Reference Table¶
| Domain | Dataset | Source | Output Column |
|---|---|---|---|
| Boundaries | Neighborhoods | ODP | — |
| Environmental | NDVI | GEE | environmental_ndvi_score |
| Environmental | Trees | PPR | environmental_tree_score |
| Environmental | Parks | OSM | environmental_park_score |
| Land Use | Healthcare | OSM | health_proximity_score |
| Land Use | Civic services | OSM | service_proximity_score |
| Mobility | Sidewalks | OSM | mobility_sidewalk_score |
| Mobility | Cartways | ODP | mobility_cartway_score |
| Mobility | Bike lanes | ODP | mobility_bike_score |
| Social | ACS 2022 | Census | social_score |
| Composite | Weighted sum | — | accessibility_score |