Mahdi Rahimi
I have been a GIS specialist for the last 15 years, seeking to broaden my GISc-related knowledge both academically and practically. Currently a postdoctoral researcher in FSS/RUG, focusing on different aspects of accessibility.
Session
AccessCity4All is a research project comparing objective, perceived, and actualized accessibility across five EU cities. Within the project, WP5 collects PPGIS data via Maptionnaire: respondents draw their perceived neighborhood on a map and mark home, work, and activity locations as points. This talk is about the R pipeline that turns those surveys into reproducible analytic datasets.
Below the survey layer everything is open: R + sf for geometry, r5r / OSRM / OpenRouteService for routing, GeoPackage and GeoJSON for interchange, and a Mistral-driven step for multilingual free-text reclassification.
What it actually taught us is what happens when around 5,000 respondents across five cities draw their own neighborhoods, around 18,000 trips need real routed distances, and free-text responses arrive in five languages. This talk walks through three defensive patterns the codebase grew over five years:
- a geometry-repair pipeline that turns freehand citizen drawings into valid analytic polygons, with home-inside verification that catches its own cache mistakes;
- a dual-mode activity-space classifier built on routed distances and pure-R clustering, with a Euclidean fallback when the routing engine is flaky;
- an audit-logged AI workflow that's defensible in peer review.
If you write geospatial R that has to survive contact with real data, this talk is for you.