GIS & cadastre

A digital cadastre is the city map used as a register: the authoritative record of who owns what, what stands where, what is taxable and what belongs to the public. It is the atlas's oldest genre — Rio's register of favelas dates from 1990 — and the one that most confuses building the map with enforcing it.

This guide covers 61 documented entries in 52 cities across 35 countries, on the register side of a line the atlas draws twice elsewhere: the digital twin guide covers the simulation model, the open data guide covers publishing under a licence. Singapore holds both — OneMap, its authoritative base map, has run since 2010, while the S$73 million Virtual Singapore twin was scaled back with no planning decision ever documented on it. A register is judged by who must consult it, not by what it renders.

What counts as a digital cadastre?

Four mechanisms recur across the documented record:

1. Taxing what is on the map. Bengaluru's GEPTIS grew from 600,000 property IDs to over 2 million, and a state review still found under half the city's properties assessed. Pune has geo-tagged since 2016 without ever publishing a recovery figure.
2. Addressing. Ghana Post cut the country into 5m x 5m squares in 2017, and a code is compulsory for Ghana Card applicants. Abidjan stencilled roughly 256,000 door numbers across 14,279 streets, each giving the distance in metres from the street's start.
3. Registering what the public owns. Bangkok logs every planting as a photographed, pinned, species-tagged object. Hyderabad's HYDRAA re-establishes each lake's Full Tank Level boundary from satellite, drone and cadastral imagery. Kigali publishes 557,949 parcels, each carrying a zone, a phase and a dated window for change.
4. Registering condition, not ownership. Frankfurt's scan cars survey road surfaces with 3D laser point clouds for a pavement management system. Los Angeles has driven and scored every street and alley since 2016 — about 22,000 miles a quarter.

Five registers worth comparing

What the evidence actually supports

Less than the parcel counts suggest: of the 61 entries, 31 state no impact figure, one is graded peer-reviewed — the GhanaPostGPS survey above — and 39 rest on official counts of what was built, not of what changed.

The sharpest finding is instead about who surveys. Some registers here were made by residents — thirteen Kibera residents with GPS handsets in 2009, hundreds of volunteers across 49 Dar es Salaam wards for Ramani Huria — and that data is the most durable material here, and the least consequential. An Overpass query on 1 August 2026 found 17,773 building outlines in two neighbourhoods of Accra's Open Cities mapping, up from 82 in June 2018 and six years after the grant ended, while the flood programme they fed had rehabilitated 0.10 km of a planned 6 km of drainage. A map anyone may copy is a weaker instrument than a register a department must consult — Map Kibera's own evaluation concedes it cannot say who uses it (more in our donor-funding essay).

Administration-run registers hold that obligation and pay elsewhere. Kigali's map gates building permits — and a peer-reviewed study attributes the city's expropriation-driven transformation to that same plan, displaced households reporting compensation too low to buy back nearby.

Where it failed or shrank

Only three of the 61 entries carry a failed or scaled-back outcome — itself a finding: registers are rarely declared dead, they go quietly obsolete. Mumbai's 2014 contract to GIS-map an undocumented water network was 18.6% executed when the deadline passed in July 2019 — the only failure here. CleanStat still scores every block, but publication thinned to one dataset a year after 2018, and while it reported streets getting cleaner the City Controller found illegal-dumping reports up 300% from 2017 to 2020. The quieter pattern carries no outcome label at all: Tunis has billed property tax online since 2009 on software released in 2008, over plain HTTP, with a news page that stops in 2012.

Questions to ask before procurement

• Who is obliged to consult it? Kigali's map gates building permits; Map Kibera's cannot say who uses it. That decides instrument or archive.
• Is the enforcement layer built? HYDRAA identified 1,025 water bodies needing demarcation, had detailed reports for 12, and its 950 perimeter cameras were still on paper.
• What does being on the map expose people to? Rio's SABREN is blunt: the boundary guiding service planning also carries the risk-area label used to justify demolitions.
• Can the public query it, or only look at it? Kigali's resident app runs on a copy with the parcel identifier stripped; Curitiba offers a curated viewer.
• Who maintains it after handover? Greater Metro Manila's hazard mapping delivered an average of 87 maps per local government, then ended; updating fell to each city.

The practical takeaway: building the register is the cheap half. Everything here that changed something had a second instrument behind the map — an audit, a permit gate, a legal boundary. Buy the obligation with the survey, or you have bought a picture of the problem.

Frequently asked