Skip to content

Indonesian street network morphology — the driving network and the walking network, measured separately for the same place.

Assumptions

Which tag values count as drivable and which as walkable is a modelling choice, not a fact. It changes every number in this product, so it is exposed as a control rather than buried as a constant.

All three mappings below are computed in full by the pipeline at build time. Switching between them shows numbers that already exist rather than recomputing in the browser — this page makes no network request at all.

What survives, and what does not

The conclusion from the tables below, computed rather than guessed. Robust: Orientation entropy — H, Circuity, Average node degree. Moderate: φ orientation-order, Dead-end proportion. Sensitive: Four-way proportion, Intersection density, Median segment length, Network length — those only mean anything stated together with their mapping, and must not be compared against figures from a study that used a different one.

The pattern makes sense. Measures that answer “which way do the streets run” survive, because adding or removing a class rarely changes the overall directions. Measures that answer “how much street is there” do not, because that is exactly what the mapping decision changes. φ falls between the two: it is derived from entropy, but the square in its formula amplifies a small movement.

The tag mapping

Tag mapping

The same mapping, drawn

The same disc, different tag mappings. The numbers in the tables above move because these are the lines coming and going. Only two exemplar sites carry these drawings: geometry is most of the derived database, and carrying it for all sixteen would multiply it for a figure shown twice.

KayutanganMalang · r = 800 m

Drive

Default · 30.2 km

Strict drive · 23.1 km

Narrow walk · 30.2 km

Walk

Default · 54.8 km

Strict drive · 54.8 km

Narrow walk · 53.4 km

BSD City — klusterTangerang Selatan · r = 800 m

Drive

Default · 26.1 km

Strict drive · 15.0 km

Narrow walk · 26.1 km

Walk

Default · 35.6 km

Strict drive · 35.6 km

Narrow walk · 28.6 km

Footway coverage per site

The headline finding depends on gang being mapped. If a kampung’s alleys are absent from OpenStreetMap its walking network collapses toward its driving network and the gap disappears — not because it is not there, but because nobody mapped it. This is a measure of the data, not of the place.

Footway coverage per site, with its confidence and density.
SiteFootway coverageConfidencem/km²
Menteng8.1%thin1522
Kota Tua10.5%thin2434
Kota Lama1.9%thin485
Kampung Bendungan Hilir18.9%moderate5989
Kampung Melayu2.6%thin700
Kampung Code19.6%moderate5150
Kampung Lette4.1%thin912
Kampung Braga35.4%good10276
Kotagede26.2%moderate7627
Kayutangan17.2%moderate4680
Pantai Indah Kapuk31.5%moderate3768
BSD City — kluster33.9%moderate6004
Alam Sutera10.5%thin1627
Gading Serpong1.4%thin334
Panakkukang2.7%thin641
Empang19.7%moderate4361
Pasar Atas21.2%moderate3675
Ubud15.8%moderate2150
Kebayoran Baru27.9%moderate7182
Cihapit25.9%moderate8487
Pathuk25.0%moderate6001
Kesepuhan20.6%moderate2988
Sanur Kaja30.5%moderate4144
Gadang25.3%moderate6988
Gapuk Utara18.0%moderate4542
Hamadi15.3%moderate1286
Cipayung36.4%good10600
Kampung Bugis27.3%moderate2109
Palangka21.9%moderate6381
Lasoani20.7%moderate3720
Penurunan15.8%moderate1977
Singkawang Tengah15.4%moderate3656
IKN — Kawasan Inti0.0%thin0