Some time ago with Missing Maps volunteers, we have started mapping in Puncak Yaja and some surrounding areas, mostly via the HOT Tasking Manager and also some via MapRoulette. When we started, I tried to get in touch with some active mappers in the region, but I had not received any feedback. Since I am in touch with colleagues who are local in the area, we still decided to start mapping. Till now this was mostly mapping buildings and highways.
However recently we have also started editing places on OSM, as I was doing yesterday. But, I was made aware via a changeset comment Changeset: 188117196 | OpenStreetMap that this might be less straightforward than I was thinking. The guideline that we at the moment more or less follow is: if there are 15 or more buildings clustered close to each other it will be place=village and if there are less than 15 buildings clustered it will be place=hamlet. Why this mapping of place=* without even knowing the name? It is a great indication for data users that there are houses and probably people living in the area.
Terkait upaya untuk menghubungi pemeta lokal: jika Anda tidak mendapat balasan langsung melalui komentar changeset atau pesan pengguna, bertanya di forum ini merupakan cara terbaik untuk mendapatkan masukan yang lebih luas dari komunitas Indonesia.
Terkait pendekatan klasifikasi untuk tag place=*:
place=village: Di Indonesia, village umumnya digunakan untuk merepresentasikan wilayah administrasi tingkat desa (setara dengan admin_level=7). Titik-titik ini biasanya diletakkan pada pusat (centroid) wilayah administratif atau di dekat kantor desanya, bukan ditentukan semata-mata berdasarkan jumlah bangunan.
place=hamlet: Permukiman yang berada di bawah lingkup desa, seperti kampung atau dusun, dipetakan menggunakan place=hamlet.
Terkait pemetaan place=* tanpa tag name=*, sepemahaman saya, umumnya tidak akan dirender pada map style utama OSM Carto, sehingga secara visual mungkin tidak menunjukkan area permukiman yang dituju. Daripada menggunakan titik place=* tanpa nama untuk menunjukkan adanya tempat tinggal, menggunakan poligon landuse=residential adalah metode yang jauh lebih disarankan. Poligon landuse=residential dengan jelas membatasi area hunian tanpa mengharuskan adanya tag nama, sekaligus tetap dirender dan menjelaskan keberadaan area permukiman.
Hi Jorieke,
Regarding reaching out to local mappers: if you don’t receive a response directly through changeset comments or user messages, posting here on the forum is definitely the best way to get broader input from the Indonesian mapping community.
Regarding your classification approach for place=*:
place=village: In Indonesia, village is typically used to represent administrative villages (Desa, equivalent to admin_level=7). These nodes are usually placed at the administrative centroid or near the local village office (kantor desa), rather than assigned solely based on building count.
place=hamlet: Settlements within a village—such as a kampung or dusun—are mapped using place=hamlet.
Regarding mapping place=* without a name=* tag, as far as I know, generally do not render on the main OSM Carto map style, so they might not visually indicate settlements as intended. Instead of using unnamed place=* nodes to show where people live, using landuse=residential polygons is usually the preferred method. A landuse=residential polygon clearly delineates the inhabited footprint without requiring a name tag, while still being rendered and showing the existence of residential areas.
Thanks again for your reply - and apologies for my late response.
I had found the Indonesian Tagging Guidelines - OpenStreetMap Wiki which I found very helpful, but I only now fully understand that you are using place=village really for places administratively represented and the settlements within villages as place=hamlet. I’ll make sure to adapt any further mapping I do or instruct remote mappers to do.
In general the rendering is not too important for our usage, we’d rather download and use the place=* data from OpenStreetMap to use it on maps to understand where people are exactly living. An example of such a reference map we make with place=* can be seen here (this example is however not in Indonesia): South Sudan - Northern Bahr El Ghazal - General Reference Map (27/02/2026) - South Sudan | ReliefWeb So as you can see it is rather the place point that we use to represent locations and not landuse=residential. I however think that we can adapt our workflow and work with landuse=residential, we will than just take the centroid of the residential area as reference points for our maps.