🗺️ GIS Risk Map — Wongwian Yai Tunnel

📊 Risk Scores by Chainage

Chainage	Score	Class	Canal	Flood	Soil

📐 Risk Formula

Canal proximity	within 50m=+3, 100m=+2, 250m=+1
Flood-risk zone	intersects=+3, within 250m=+2, 500m=+1
Elevation (low point)	<1m=+2, <3m=+1
Building density	dense=+2, medium=+1
Soil/geology	marine clay=+2, moderate=+1
Rainfall	heavy=+2, moderate=+1
Total	0-3=LOW · 4-7=MEDIUM · 8+=HIGH

Source: User's QGIS document — Wongwian Yai tunnel risk formula.

🔄 Data Pipeline

📂 OSM (Geofabrik)
canals, roads, buildings

→

🗺️ QGIS
or GeoLibre
buffer, intersect, join

→

🐍 risk_scorer.py
geopandas+shapely
compute scores

→

📊 GeoJSON
risk segments

→

🗺️ MapLibre
this dashboard

→

🔮 Anomaly Detector
real-time alerts

Open data identifies WHERE risk is highest. Engineering telemetry (MRTA piezometers, TBM, pumps) detects WHEN an incident starts. Two separate pipelines, one dashboard.

🛠️ Tools & Files

🐍 risk_scorer.py

Python CLI — reads GeoJSON layers, computes 6-factor risk per 100m chainage, outputs GeoJSON.

python3 gis/risk_scorer.py --sample --pretty
python3 gis/risk_scorer.py --alignment alignment.geojson --canals canals.geojson --flood flood_zones.geojson --soil soil_zones.geojson --output output.geojson

/root/sinkalert/gis/risk_scorer.py

🗺️ QGIS + MCP

PyQGIS script + AI prompt for QGIS-MCP. Loads alignment, OSM layers, buffers, risk scores, exports PDF map + CSV table.

/root/sinkalert/gis/qgis_tunnel_risk.py
/root/sinkalert/gis/qgis_mcp_prompt.txt

📂 GeoJSON Outputs

Pre-computed risk segments + sample input layers. Load into QGIS, GeoLibre, or MapLibre.

wongwianyai_risk_segments.geojson
alignment.geojson · canals.geojson
flood_zones.geojson · soil_zones.geojson

⚠️ Honest Limitations

Sample data is approximate. Tunnel alignment digitized from MRTA route map — not engineering CAD. Real alignment from MRTA would be more accurate.
OSM data is community-sourced. Canals and buildings may be incomplete or outdated. Geofabrik Thailand extract from July 2026.
BMA flood data not loaded live. ArcGIS REST endpoint at cpudapp.bangkok.go.th — requires browser/session auth. QGIS can add it as ArcGIS REST layer when on the same network.
This is static risk. Dynamic risk requires real-time sensor data (piezometers, TBM telemetry, pump rates) — not publicly available.
30m DEM is coarse for Bangkok. Bangkok elevation range is only 1-5m. 30m SRTM cannot capture micro-topography. LiDAR DEM would be needed for engineering precision.
Soil data is regional. SoilGrids gives 250m resolution soil proxy. Borehole logs from MRTA/contractors would provide exact clay thickness and SPT-N values.

🗺️ Wongwian Yai Tunnel Risk Map

📍 Risk Map — Purple Line South (Wongwian Yai Section)