The challenge
Cadastral data, building permits, satellite imagery and mobility information usually live in systems that do not talk to each other, leaving planning teams unable to anticipate problems.
The solution
- PostGIS to model and query geospatial data.
- Supervised and unsupervised models to detect growth and risk patterns.
- Integration of historical series, satellite imagery and mobility data.
- Visualisation with Mapbox GL and interactive layers.
- Dashboards for technical teams and authorities, with a REST API and webhooks for municipal systems.
Results
Analyses that used to take weeks now take hours, with early identification of critical areas on a single, shared data base.
| Before | After | |
|---|---|---|
| Analysis time | Weeks | Hours |
| Data sources | Scattered | Unified in PostGIS |
| Decisions | Reactive | With risk anticipation |
Use cases
Municipalities, regional governments, urban planning teams and housing ministries.
In depth
Sources that were never designed together
A building permit, a cadastral record and a satellite image describe different things at different times. Urban A.I records lineage, scale and update dates so comparisons can be reconstructed rather than simply viewed.
An architecture for scenario research
The geospatial layer organises geometries and time series; query services feed maps and dashboards; modelling experiments remain separate from operational data. Versioning allows teams to compare and retire an experiment safely.
Human decisions and model limits
A risk map is not an instruction. Technical teams must compare model signals with field work, policy changes and residents’ knowledge. False positives need documenting alongside useful findings.
Interfaces for planning
Planners need comparisons over time, analysts need definitions and outlier review, and leadership needs a concise account of trends and open questions. Exports must include dates, coverage and method.
From urbanism to new questions
Territorial methods can inform new work, including mining, but operational safety and community relationships require their own research. Mining is a developing line, not a completed Urban A.I deployment.