Today a found out, that exported archives for countries (Turkey, Czech republic), that had a lot of data (5-10 MB zips and more) last autumn, became 10 times smaller now.
Archives for smaller countries like Armenia collapsed from reasonable 300 - 350 Kb to tiny 32 Kb.
The root cause seems to be a new (unannounced?) policy for data export, which is expressed in a small note, visible just for authorised users at Data Downloads - OpenCelliD - Largest Open Database of Cell Towers & Geolocation - by Unwired Labs
“Exports… …include cell towers observed in the last 18 months”
Such a policy can be pretty effective and reasonable for areas, where you have regular data update – SF, LA, NY, etc, but seems to be catastrophic for rarely visited zones where there are just unique data.
As for me, I started contributing to OpenCelliD about a year ago to build a map of cell coverage in rural and mountain regions of one of the Central Asia countries to provide tourists with information: in case of emergency, you can find mobil network there or there.
Even for many main roads of this country i was the first to collect public cell coverage data.
I do not see a lot of followers around and i do not expect someone to reobserve my data in mountainious areas, where they are especially valuable.
This policy means all my work would be lost in 12 or 18 monthes.
My friends are visiting Mongolia these days. I wanted to say them “for cheap communications buy a sim card from provider X”, but Mongolia seems to be almost empty.
Empty or just not observed in last 18 monthes?
I’d like you to rethink your data-export policy once again, and to keep exporting ancient data for areas, you don’t have modern observations.
Scincerely yours, Serge.
P.S. Yes, i do understand, that 18 month threshold is an easy and cheap way to eleminate data for dead providers, decomissioned cells etc.