Ten regions, one method, proven in Bontang first.

Presidential Instruction No. 4/2022 tasked every governor with coordinating and updating extreme-poverty target data toward a 0% national rate by 2024. Kalimantan Timur has the 10th-lowest poverty rate in Indonesia, but 6.31% of its population, up 0.04 points from the prior period, still meant close to a quarter-million individuals flagged extremely poor across the province’s ten kabupaten/kota.

P3KE, Indonesia’s national extreme-poverty registry, already holds per-individual, address-level records for every province. What it doesn’t provide is any built-in way to see where those households cluster. Kutai Kartanegara alone carries 65,380 flagged individuals, nine times Bontang’s 7,297, a gap invisible in a spreadsheet, and the whole reason this mapping study exists.

Sasaran 1 digitized every P3KE individual as a point on the road network, joined to their household attributes, then ran kernel density estimation to turn scattered points into a concentration surface. That is the same technique the Cikarang industrial-settlement study uses to turn scattered building footprints into a concentration surface, applied here to people instead of buildings. Sasaran 2 ran a separate characteristic-cluster analysis across 19 P3KE indicators (gender, work, education, housing materials, utilities, aid-program enrollment, and child stunting risk) to explain what a hotspot is actually made of.

By this report’s cut-off, Bontang was the only region at 100% on both the concentration map and the characteristic clustering. Mahakam Ulu and Penajam Paser Utara had only the hotspot map done; Kutai Barat was 75% through it; the remaining six (including Samarinda and Kutai Kartanegara, the two largest poor populations in the province) hadn’t started either track yet. Bontang’s comparatively small caseload, second-smallest of the ten, made it the tractable place to prove the method first.

Bontang’s two output maps (raw point distribution and kernel-density concentration) show the same households from two angles: one where poverty reads as scattered dots across the city, the other where kernel density collapses those dots into a small number of dense pockets. That distinction is the entire point of the method: a scattered spread and a concentrated pocket carrying the same headcount call for different interventions.

Of Bontang’s dozens of kelurahan profiled under Sasaran 2, Tanjung Laut Indah (Kecamatan Bontang Selatan, 590 KK sampled) shows what the 19-indicator layer adds. That is the same indicator count, coincidentally, that scores each kecamatan in the Jabung sub-region hierarchy study: half the sampled households rent rather than own, 69% report no savings, valuables, or livestock to fall back on, and 60% of children sit in the middle stunting-risk band. None of this shows up in a raw headcount. It’s exactly the texture a program needs to decide whether to lead with housing, livelihood, or health support.

The same P3KE pipeline that mapped Bontang now has a working, repeatable shape. The province’s task is running it nine more times. The study also became more than a static report: a live web GIS tool now lets planners query individual-level characteristics by location directly, the same per-individual data this analysis draws on, without waiting for the next paparan deck. That is the same shift, from one-off report to a live tool a planning team keeps using, that closes South Papua’s RPPLH.