Large numbers of broken boundaries

The Data Working Group have received complaints from Iranian OSM users about a number of users adding boundaries in Iran. These seem to be added as “source=Snapp” or “source=snapp”. Snapp appears to be a private hire company in Iran; it is unclear how it may be the source of boundary data. As well as the unclear source, these boundaries have often been added with errors (e.g. not closed) which has resulted in other users having to remove them again.

Attempts have been made to try and contact these users. No (or almost no) responses have been received.

We’d be grateful if anyone who can influence these users could try and get them to engage with the rest of the OSM community and discuss the work being carried out, the licences under which the data is made available and the methodology for the import.

Best Regards,

Andy Townsend, on behalf of OSM’s Data Working Group.

hi Andy , we managed to contact some of them and they told us they currently stopped importing this boundary to fix issues was made by theme. user Amin Saleh is one of the Snapp good guys which we are in contact with in our Telegram group chat. the errors are getting fixed slowly by our community so i think they should not get removed.
he wrote in another separated topic :

Good news - thanks.

I worked a bit on Arak and Bushehr cities and fixed the errors on other elements which were came up as a side effect of these edits. I also unglued the new boundaries from other elements. The borders of neighbourhoods, themselves, still need attention. I’m waiting for some good source/information to work on them. I know some users are/were working on some other places…

What actually is the source of the new boundaries, and do they form part of an administrative hierarchy, or are they instead something like “place=neighbourhood”?

Yes, they are part of administrative hierarchy in Iran which go under municipalities. We call them “محله” (Mahalleh) and they have known borders specified by government based on local characteristics. But the work that was started by Snapp in some cases was a rough estimation by the help of local people, because of limited access to the data. In some cases it was a very bad estimation (e.g. drawing a boundary around a famous street). Many of them has been removed, and also they agreed on this to not do this way. In some other cases it’s good (like Tehran). Also we found some better information (near to reality) for other cities like Mashhad, that they said will use it.

Imho this estimations is much better than placing a single neighbourhood point for them, which leads to a bad addressing.