New Urban Mapping datasets: Tashkent and Ulanbaator

After the long break we started publishing more datasets under the Open Urban Mapping project.
Initially the project was aimed at demonstrating “AI-Mapping” technology of Deep learning models powered by the Mapflow platform to extract millions of building footprints. As far as Microsoft Building footprints continues covering territories all over the World it comes to little sense to develop an open data alternative to this gigantic datasets. 👉 Therefore we reviewed the Urbam Mapping project concept and dedicated it to the much less in numbers but more sophisticated and reach datasets related to particular cities rather than the whole countries. The datasets keep on publishing under the Open (OdbL) license.

There are two of them on the current agenda we started with. Each dataset composed of different additional layers to extend the potential scope of this data application for the urban analysis and research.


Tashkent (capital of Uzbeklistan) dataset is enriched with the building heights estimates. Althoug the accuracy is not so high, the reference data can be used to improve the model’s prediction and get it more solid. There are also the two of building footprints layers — whether the data is merged with Openstreetmap buildings or not.

In the last years Tashkent is actively developing with new building constructions — that’s why we added the polygons of construction sites, around 500 were detected. Note, the imagery we had to use due to the compatibility with the open license is not so recent and these areas of construction are changing fast.

Tashkent patterns on an overview


Ulanbaator is the capital and the largest city of Mongolia. The interesting fact about it that more than 1/3 of all constructions in the city are yurts. They used to call them “Ger” in Mongolian.

Modern urban planning began in the 1950s, with most of the old Ger districts replaced by Soviet-style flats. In 1990, Ulaanbaatar was a major site of demonstrations that led to Mongolia’s transition to democracy and a market economy. Since 1990, an influx of migrants from the rest of the country has led to an explosive growth in its population, a major portion of which live in Ger districts, which has led to harmful air pollution in winter.

Ulanbaator yurts and new buildings (source)

To check this statistics we took the data that was published by Alexander, CV Engineer at Geoalert. He published the dataset and training pipelines on and posted about the model development in his blog:

There is a special tag in Openstreetmap “Tag: building=ger” and many yurts are already mapped. But I have done much more, and covered whole Ulaanbaatar, not just downtown — it is more, than 4 700

In the table below you see the number of features related to the usual buildings and to the yurts which makes about 1/4 of all constructions.

The geodata layers for Ulanbaator

Data accessibility

One more important thing is to provide these datasets in GPKG format and via #API for more convenient and interoperable use by GIS users and developers.

👉 Tell us what do you think and what cities you’d prefer to get published next.




Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store

We apply Machine learning to automated analysis over Earth observation data