Methodology (Brief Overview)

This project constructs a multidimensional Neighborhood Accessibility Index for Philadelphia by combining four analytic domains:

• Mobility (sidewalks, bike lanes, curb/cartway structure)
• Land Use & Amenities (health and civic service proximity)
• Environmental Quality (NDVI, parks, tree canopy)
• Social Conditions (ACS demographic indicators)

Because each domain required different data sources and spatial operations, the full methodological details are documented inside each domain’s dedicated markdown page:

• Mobility Score
• Environmental Score
• Land Use Score
• Social Score
• Accessibility Score

All components follow a consistent workflow:

---

1. Data Processing

Spatial data from OSM, OpenDataPhilly, Census ACS, and Google Earth Engine were cleaned, reprojected, and clipped to the Philadelphia boundary.
OSM-derived features (sidewalks, parks, amenities) were retrieved via OSMnx.

---

2. Metric Construction

Each domain required a specific computation method:

• Distances: KDTree nearest-facility calculations
• Environmental Raster Stats: NDVI zonal statistics
• Line Density Measures: sidewalk, bike lane, and cartway length per neighborhood
• ACS-derived rates: disability, elderly, commuting behavior

---

3. Normalization

All variables were scaled to a 0–1 MinMax range to allow comparability:

\[ \text{score} = \frac{x - x_{\min}}{x_{\max} - x_{\min}} \]

Some variables were inverted so that:

• higher = more accessible

---

4. Aggregation

Metrics computed at the tract level were aggregated to neighborhoods using spatial joins and mean statistics.

---

5. Composite Scoring

Each domain score was normalized, combined internally, and then incorporated into the final weighted index:

\[ \text{AccessibilityScore} = 0.4\,\text{Mobility} + 0.3\,\text{LandUse} + 0.2\,\text{Environmental} + 0.1\,\text{Social} \]

---

This page provides the conceptual workflow.
See the analysis sections for fully detailed formulas, maps, and interpretation.