Automatically Correcting Spatial Validations

Cover

Inspiration

We were inspired by the need for more accurate location data in cities like Mexico City, where incorrect POI placements can disrupt navigation, logistics, and urban planning.

What it does

Our solution detects and categorizes spatial errors in Points of Interest (POIs), such as incorrect placement, outdated data, or wrong-side-of-the-road associations, using satellite imagery and road network data.

How we built it

We used Python with GeoPandas and Folium to process and visualize POI and street datasets. We integrated HERE Maps API to retrieve satellite tiles and created a rule-based pipeline to classify and flag spatial inconsistencies.

Challenges we ran into

Aligning different geospatial formats and coordinate systems.
Matching POIs with correct road segments in dense urban areas.
Handling missing or inconsistent data across multiple sources.

Accomplishments that we're proud of

Built an automated validation pipeline that works at scale.
Achieved reliable classification of four POI error types.
Created an interactive map to visually inspect and validate outputs.

What we learned

We deepened our understanding of spatial data alignment, geocoding complexities, and the value of combining visual context (satellite) with geometric analysis for accurate validation.