Metro-Mapping

Drawing Philippine metro areas from open data

City limits often miss the place people actually experience: some Philippine cities contain large rural areas, while other urban regions continue across several local-government boundaries. This project estimates the connected built-up footprint around each city's centre.

1

Divide the region

The city boundary and a 12 km surrounding area are divided into H3 hexagons. Water cells are removed.

2

Identify urban cells

A cell enters the extent when it holds at least 3 distinct OpenStreetMap establishments, or at least 0.75 km of physical road backed by nearby establishments and a WorldPop density of 300 people per km², or belongs to a high-density centre of at least 50,000 people. The core applies the same rule inside the extent with a 1,500 per km² floor. Bars were calibrated on five cities against the GHS-SMOD grid, then frozen.

3

Connect and time

Both tiers keep the urban cells connected to the detected city centre; short water crossings can be bridged, ordinary rural gaps stay outside. Because Philippine commuting runs on roads, every cell also gets its free-flow road travel time to the centre, shown as 45- and 60-minute reach.

2.4%Puerto Princesa extent / PSA land area
334%Iloilo extent / PSA land area
0.56Held-out core IoU vs GHS-SMOD dense classes (12 cities)

How to read the map

The red dashed line is the settled extent and the solid dark line is the dense core; the core holds most of the population on a fraction of the area. Both are exploratory built-up activity footprints, useful for comparing the urban extent with administrative limits, but it is not yet a commuting-based official metropolitan boundary. Area ratios use declared PSA land areas rather than the OpenStreetMap polygon, and every footprint carries a stable core that survives all nine tested threshold settings. Results that reach the study edge, rest on a proxy boundary or lack a qualifying population centre are flagged in the app.

Population counts are summed from WorldPop 2020. OpenStreetMap supplies roads, establishments, administrative boundaries and water. Night lights remain a diagnostic layer. Commercial land-price estimates are retained as a separate experiment and are not part of the metro boundary.

Explore 37 cities View source