radiocells-scanner-android
radiocells-scanner-android copied to clipboard
openbmap
Cooperate with OpenCellID & Mozilla location database ?
Opencellid currently maints open database of cell towers and Mozilla has cell & WIFI database. They started cooperating some time ago - i.e. sharing collected information thus benefiting the whole community. I am currently collecting cell towers for opencell and for radiocells.org.
But I feel it doesn't really serve anyone well to have several open tower databases. Have you considered cooperation with above mentioned parties. It is waste of time and resources to collect same gsm tower and wifi hotspots for several parties.
A critical issue is the degree to which raw data is available. If there's a way to synchronize multiple ways, and have the synchronized data available under the same terms radiocells has now, that's probably a good thing to do. That might be doable with cell data, but seems not currently possible with wifi.
Another issue is having a way to merge public files for a combined NLP backend, or a web service.
Have you considered cooperation with above mentioned parties.
It's a small community, so we've been in touch some time ago. The issue is not on the technical side, but there's the raw data access issue, which @gdt already mentioned: Raw data access is essential for all in-depth analysis, especially for any scientific research. We currently offer this data whereas access at the other projects is limited:
to my knownledge Mozilla doesn't offer any cell raw data access at all, opencell id has a raw cell data format, but it's IMHO difficult to work with (I think I had to parse one single very huge CSV file with all measurements with brought my hardware to it's limits, situation may have changed in the meantime though).
For the wifi part the situation is even worse: Mozilla won't give you free access to wifi data at all. The only way to access their wifi data is using their 'protected' location APIs, meaning you've to prove that you're physically at the place before getting any API results: This is perfectly reasonable and OK for geolocation purposes, but it's way too little for any kind of research.
I would definetely support any form of cooperation or synchronization, but I would like to keep the radiocells.org raw data access for the community at all costs too.
Regarding the specific geolocation part, you might be interested in our friend Alex's meta geolocation database available at https://www.mylnikov.org/ Alex is currently using radiocellls.org and openwifi.su data and shows a way to aggregate open source geolocation data. This might be a starting point for a unified open source geolocation database too.
On the issue with the csv file from opencellid; this causes problems for me too. From what I recall the client downloads a several hundred MB file and then looks through for the local MNC/MCC. It takes a long time on my Samsung, but there is a script by lacells which generates it on a fast device which one can then transfer to the phone. I have a weekly cron job doing this on my server for NZ, then I wget it to my phone every so often; much quicker and easier.
OpenCellID is closing (see the referenced issue). Years of contributors' work might end being flushed down the toilet. Any way to fuse their database with yours?
OpenCellID is closing (see the referenced issue). Years of contributors' work might end being flushed down the toilet. Any way to fuse their database with yours?
Just fired up a test import for the aggregated opencellid tower information.. Seems pretty compatible.. As soon as things are ready I'll enable a testing layer on the map
Regarding the opencellid raw data we still have to find a solution as this a gigabytes of CSV to process.
@takiainen We're currently discussing ways/options on this issue via email.. If you (or anybody else) like to join our discussion, drop me your mail address to wish7code @ gmail . I'll then include you in our discussion..
Sorry, missed out the details. Your guess is correct:
It's a test import of the OCID database. It's roughly 35 mio. aggregated OCID cell tower locations. The good news from the first tests: their aggregated cell tower information is pretty compatible with the radiocells.org format. Raw data may be a bit harder due to the massive dataset though. Let's see..
Strange thing: US, Russia and China have disappeared on the opencellid homepage... Guess we need a backup quickly..
It seems they are not closing after all. Another company has taken over and is currently transferring the database on their servers. They said they will publish their roadmap soon.
Anyways, hope to see ocid data merged with radiocells.org soon.