Breakout Session
summary
Geocoding and reverse geocoding run inside PostgreSQL by calling web services and caching results; modules for OpenStreetMap (nominatim) and Esri ArcGIS (Ares) return coordinates per-address via SQL functions. On test sets Ares resolves ~100% US and ~84% global (165/196 non‑US), nominatim resolves ~60–70% US and ~36% global (71/196); overlap is 68 addresses. Inter‑service differences range from ~6.7 m to >6 km (mean ~540 m). Implementation works with ETL caching and parallel Greenplum calls but misses China/Korea and needs better error handling.
AI-written summary · about
chapters (16)
- 0:00Address geocoding with OpenStreetMap
- 1:05Why address resolution matters
- 3:02Address format challenges
- 8:24Zip code and variation problems
- 13:01Geocoding services compared
- 16:35Demo: Address to coordinates
- 20:21Demo: Bulk address resolution
- 24:29Ares vs OpenStreetMap results
- 27:23Geocoded data applications
- 31:55Geocoding performance findings
- 39:06Technical limitations
- 40:20Future improvements
- 43:37Caching and error handling
- 47:03Local OSM data with Nominatim
- 51:19Google Maps restrictions
- 52:10Conclusion and Q&A
related talks
20mSolving a Secure Geocoding Problem (That Hardly Anybody Has)
50m–Mom, can we have G**gle Maps? –We have G**gle Maps at home
02h 44mBuilding Geographical APIs with Open Data, PostgreSQL & PostGIS
46mGetting Down To Earth: Geospatial Analysis With Rails
40mGIS capability of PostgreSql using POSTGIS & OSM DATA
16mSelf hosted GIS with open geo data - Real-world experience
16mMapping the Census with PostGIS, Python, and...
30mGeospatial analysis made easy with PostGIS and Geoalchemy