Skip to main content

Urban Planning: Properties Near Transit

Find all residential parcels within walking distance of subway stations for transit-oriented development analysis.

Scenario Overview​

Goal: Identify properties within 500 meters of subway stations to assess transit-oriented development opportunities.

Real-World Application:

  • Urban planning departments evaluating development zones
  • Real estate developers finding transit-accessible properties
  • Policy makers assessing transit equity and coverage
  • Environmental planners reducing car dependency

Estimated Time: 10 minutes

Difficulty: ⭐⭐ Intermediate

Prerequisites​

Required Data​

  1. Parcels Layer (polygons)

    • Residential property boundaries
    • Must include land use or zoning attributes
    • Recommended: 1,000+ features for realistic analysis

  2. Transit Stations Layer (points)

    • Subway/metro station locations
    • Includes station names
    • Covers your study area

Sample Data Sources​

Option 1: OpenStreetMap (Free)

# Use QGIS QuickOSM plugin
1. Vector → QuickOSM → Quick Query
2. Key: "railway", Value: "station"
3. Select your city/region
4. Download points

Option 2: Municipal Open Data

  • Check your city's open data portal
  • Look for "parcels", "cadastre", or "property" datasets
  • Transit data usually under "transportation"

System Requirements​

  • Recommended Backend: PostgreSQL (for 50k+ parcels)
  • Alternative: Spatialite (for <50k parcels)
  • CRS: Any (FilterMate handles reprojection automatically)

Step-by-Step Instructions​

Step 1: Load Your Data​

  1. Open QGIS and create a new project
  2. Load the parcels layer (drag & drop or Layer → Add Layer)
  3. Load the transit_stations layer
  4. Verify both layers display correctly on the map
CRS Check

Different CRS? No problem! FilterMate automatically reprojects layers during spatial operations. You'll see a 🔄 indicator when reprojection occurs.

Step 2: Open FilterMate​

  1. Click the FilterMate icon in the toolbar
  2. Or: Vector → FilterMate
  3. The panel docks on the right side

What you should see:

  • Three tabs: FILTERING / EXPLORING / EXPORTING
  • Layer selector at the top
  • Empty expression builder

Step 3: Configure the Filter​

3.1 Select Target Layer​

  1. In the Layer Selection dropdown (top of panel)
  2. Check parcels layer
  3. Notice the backend indicator (PostgreSQL⚡ / Spatialite / OGR)

Layer Info Display:

Provider: postgresql (PostgreSQL)
Features: 125,347
CRS: EPSG:2154 (Lambert 93)
Primary Key: gid
Backend Performance

If you see "OGR" for large parcel datasets, consider migrating to PostgreSQL for 10-50× faster performance. See Backend Guide.

3.2 Add Attribute Filter (Optional)​

Filter to residential parcels only:

  1. In the Expression Builder section
  2. Click the Fields dropdown to see available attributes
  3. Enter this expression:
land_use = 'residential'
-- OR if using zoning codes:
zoning LIKE 'R-%'
-- OR multiple residential types:
land_use IN ('residential', 'mixed-use', 'multi-family')
  1. Wait for the green checkmark (✓) - indicates valid syntax

Expression Explanation:

  • land_use = 'residential' - Exact match on land use field
  • LIKE 'R-%' - Pattern matching for residential zoning codes (R-1, R-2, etc.)
  • IN (...) - Multiple allowed values
No Residential Field?

If your data doesn't have land use, skip this step. The spatial filter will work on all parcels.

3.3 Configure Geometric Filter​

Now add the spatial component - proximity to transit:

  1. Scroll down to the Geometric Filter section
  2. Click to expand if collapsed

Reference Layer: 3. Select transit_stations from the dropdown 4. The reference layer icon appears: 🚉

Spatial Predicate: 5. Select "Intersects" from predicate dropdown

  • (We'll add buffer distance, so intersects = "touches buffer")

Buffer Distance: 6. Enter 500 in the distance field 7. Select meters as the unit 8. Leave buffer type as Round (Planar) for urban areas

Your Configuration Should Look Like:

Reference Layer: transit_stations
Spatial Predicate: Intersects
Buffer Distance: 500 meters
Buffer Type: Round (Planar)
Geographic CRS Auto-Conversion

If your layers use geographic coordinates (EPSG:4326), FilterMate automatically converts to EPSG:3857 for accurate metric buffers. You'll see: 🌍 indicator in logs.

Step 4: Apply the Filter​

  1. Click the Apply Filter button (big button at bottom)
  2. FilterMate executes the spatial query

What Happens:

-- Creates optimized materialized view
CREATE MATERIALIZED VIEW temp_filter AS
SELECT p.*
FROM parcels p
WHERE p.land_use = 'residential'
  AND EXISTS (
    SELECT 1 FROM transit_stations s
    WHERE ST_DWithin(
      p.geom::geography,
      s.geom::geography,
      500
    )
  );

CREATE INDEX idx_temp_geom 
  ON temp_filter USING GIST(geom);

⚡ Performance: 0.3-2 seconds for 100k+ parcels

Step 5: Review Results​

Map View:

  • Filtered parcels are highlighted on the map
  • Non-matching parcels are hidden (or greyed out)
  • Count displayed in FilterMate panel: Found: 3,247 features

Verify Results:

  1. Zoom to a transit station
  2. Select one filtered parcel
  3. Use Measure Tool to verify it's within 500m of station

Expected Results:

  • Urban cores: High density of filtered parcels
  • Suburban areas: Sparse parcels near stations
  • Rural areas: Very few or no results

Step 6: Analyze & Export​

Option A: Quick Statistics​

  1. Right-click filtered layer
  2. Properties → Information
  3. View feature count and extent

Option B: Export for Reporting​

  1. Switch to EXPORTING tab in FilterMate

  2. Select filtered parcels layer

  3. Choose output format:

    • GeoPackage (.gpkg) - Best for QGIS
    • GeoJSON - For web mapping
    • Shapefile - For legacy systems
    • PostGIS - Back to database

  4. Optional: Transform CRS (e.g., WGS84 for web)

  5. Click Export

Export Settings Example:

Layer: parcels (filtered)
Format: GeoPackage
Output CRS: EPSG:4326 (WGS84)
Filename: transit_accessible_parcels.gpkg

Understanding the Results​

Interpreting Feature Counts​

Example Results:

Total parcels: 125,347
Residential parcels: 87,420 (70%)
Transit-accessible residential: 3,247 (3.7% of residential)

What This Means:

  • Only 3.7% of residential parcels are transit-accessible
  • Opportunity for transit-oriented development
  • Most residents depend on cars (equity concern)

Spatial Patterns​

Look for:

  • Clusters around major transit hubs → High-density zones
  • Gaps between stations → Potential infill development
  • Isolated parcels → Transit deserts requiring service expansion

Best Practices​

Performance Optimization​

✅ Use PostgreSQL for parcel datasets >50k` features

  • 10-50× faster than OGR backend
  • Sub-second query times even on 500k+ parcels

✅ Filter by attribute first if possible

  • land_use = 'residential' reduces spatial query scope
  • 30-50% performance improvement

✅ Buffer Distance Units

  • Use meters for urban analysis (consistent worldwide)
  • Avoid degrees for distance-based queries (inaccurate)

Accuracy Considerations​

⚠️ Buffer Type Selection:

  • Round (Planar): Fast, accurate for small areas (<10km)
  • Round (Geodesic): More accurate for large regions
  • Square: Computational optimization (rarely needed)

⚠️ CRS Choice:

  • Local projected CRS (e.g., State Plane, UTM) - Best accuracy
  • Web Mercator (EPSG:3857) - Good for worldwide analysis
  • WGS84 (EPSG:4326) - Auto-converted by FilterMate ✓

Data Quality​

🔍 Check for:

  • Overlapping parcels - Can inflate counts
  • Missing geometries - Use "Check Geometries" tool
  • Outdated transit data - Verify station operational status

Common Issues & Solutions​

Issue 1: No Results Found​

Symptoms: Filter returns 0 features, but you expect matches.

Possible Causes:

  1. ❌ Buffer distance too small (try 1000m)
  2. ❌ Wrong attribute value (check land_use field values)
  3. ❌ Layers don't overlap geographically
  4. ❌ CRS mismatch (though FilterMate handles this)

Debug Steps:

-- Test 1: Remove attribute filter
-- Just run spatial query on all parcels

-- Test 2: Increase buffer distance
-- Try 1000 or 2000 meters

-- Test 3: Reverse query
-- Filter stations within parcels (should always return results)

Issue 2: Slow Performance (>30` seconds)​

Cause: Large dataset with OGR backend.

Solutions:

  1. ✅ Install PostgreSQL + PostGIS
  2. ✅ Load data into PostgreSQL database
  3. ✅ Use PostgreSQL layer in QGIS
  4. ✅ Re-run filter (expect 10-50× speedup)

Quick PostgreSQL Setup:

# Install psycopg2 for QGIS Python
pip install psycopg2-binary

# Or in OSGeo4W Shell (Windows):
py3_env
pip install psycopg2-binary

Issue 3: Results Look Wrong​

Symptoms: Parcels far from stations are included.

Possible Causes:

  1. ❌ Buffer distance in wrong units (degrees instead of meters)
  2. ❌ "Contains" predicate instead of "Intersects"
  3. ❌ Reference layer is wrong (roads instead of stations)

Verification:

  1. Use QGIS Measure Tool
  2. Measure distance from filtered parcel to nearest station
  3. Should be ≤ 500 meters

Next Steps​

Advanced Techniques​

Graduated Buffers: Run multiple filters with different distances (250m, 500m, 1000m) to create walkability zones.

Combine with Demographics: Join census data to estimate transit-accessible population.

Time-Based Analysis: Use historical data to track transit-oriented development over time.

Summary​

You've Learned:

  • ✅ Combined attribute and geometric filtering
  • ✅ Buffer operations with distance parameters
  • ✅ Spatial predicate selection (Intersects)
  • ✅ Backend performance optimization
  • ✅ Result export and CRS transformation

Key Takeaways:

  • FilterMate handles CRS reprojection automatically
  • PostgreSQL backend provides best performance for large datasets
  • 500m is typical "walking distance" for urban planning
  • Always verify results with manual measurement sampling

Time Saved:

  • Manual selection: ~2 hours
  • Processing Toolbox (multi-step): ~20 minutes
  • FilterMate workflow: ~10 minutes ⚡

Need help? Check the Troubleshooting Guide or ask on GitHub Discussions.