[Imports] Looking for POI datasets

Jo winfixit at gmail.com
Thu Oct 22 08:50:32 UTC 2015

URbis opened a dataset with a lot of geodata several years ago for the
region of Brussels.



2015-10-22 8:35 GMT+02:00 Sander Deryckere <sanderd17 at gmail.com>:

> 2015-10-22 1:21 GMT+02:00 Elliott Plack <elliott.plack at gmail.com>:
>> Sander,
>> Cool idea! I have a number of datasets in mind, but the two most foremost
>> are ones created by local governments to inventory all their facilities
>> (schools, police stations, fire depts). A lot of that gets added to the map
>> already, so it could provide for some good QC.
>> If the DB gets split per feature type, that would be nice datasets to add.
>> The other is a gov't maintained POI dataset used in 911 dispatch (when a
>> caller doesn't know the address of the POI they're in). I would one day
>> like to do some kind of ETL to get some of that in OSM, but I also like to
>> survey these things.
>> That's what the tool is meant for, I hope it gains some popularity, and
> people use it to map in their area (or in an area they surveyed).
>> Let me know if you want them to test.
>> Elliott
> Sure, I would like to test these.
> 2015-10-21 17:16 GMT+02:00 Michael Reichert <nakaner at gmx.net>:
>> Hi Sander,
>> Am 2015-10-21 um 16:56 schrieb Sander Deryckere:
>> > I'm developing a tool called POI Importer (
>> > http://wiki.openstreetmap.org/wiki/POI_Importer ). It's a combination
>> of QA
>> > and community-powered import.
>> >
>> > The aim is that different databases of POIs will be uploaded to the
>> tool,
>> > then the tool will compare it with the existing OSM data, and allows
>> users
>> > to import the POI one by one in their area.
>> >
>> > I tried to combine the following:
>> >
>> >    - Local verification: by showing the points on a map (instead of
>> picking
>> >    a random one like maproulette), the users will most likely stay
>> close to
>> >    the area they know well, which lowers the chance of copying errors.
>> >    - Support for multiple tags at once: many POI datasets contain
>> multiple
>> >    tags. Think about name, operator, opening hours, contact info, ...
>> Some of
>> >    that data might be already in OSM, some not.
>> >    - Data is easy to update: The comparison with OSM happens live
>> (thanks
>> >    to Overpass API), so it's just a matter of updating the 3rd party
>> data when
>> >    there's a new dataset available
>> Your tool is similar (but different) to the POIchecker by Sozialhelden
>> (the organization behind Wheelmap)
>> https://github.com/sozialhelden/poichecker
>> http://poichecker.de/
>> The aim does look very similar, but I have a feeling that POIchecker is a
> bit harder to use (I don't really see how to use the app, but then I'm also
> the most biased person in here).
>> > So far, I've been developing with the dataset of our PT bus company (see
>> > http://poi-importer.github.io/#map=14/50.9343/4.0518&datasets=BE_dl )
>> this
>> > data is rather dense, which was ideal to test the speed.
>> >
>> > Now I'm looking to find other datasets. Thinking of big chains with a
>> > number of shop locations, or certain umbrella groups representing a
>> number
>> > of small enterprises. These companies will usually have no (or fewer)
>> > problems with giving that data away for free (it's free publicity for
>> > them). So I think we should be able to find some usable databases.
>> >
>> > Does anyone currently have POI datasets that could be used?
>> You might Sozialhelden if they have some data for private testing. You
>> may also ask the people which proposed and performed an Aral import in
>> Germany a few months ago. (You might use the OSM data before the import,
>> otherwise your test may end with too much good results)
>> I'll look into that. The Aral import indeed also looks promising.
>> What about a look in at an open-data website of a town/country of your
>> choice? They also have data which is mapped at OSM.
> Hehe, our towns don't do open-data (they might provide data, but never
> give a license, so all IP rights are implied). Besides, our towns are quite
> small, so their datasets are small. It's easier to work with bigger
> datasets that have many POI in a standardised format. Our region (Flanders)
> does have an open dataset with all private businesses (at least they call
> it so), but their license requires registration to use the data (so also
> forbids sharing). That doesn't really help us any further.
> The agencies responsible for GIS do realise the importance of an open
> license, and are opening multiple datasets for us now (they have opened a
> database with all addresses, and will open a bigger database with building
> and street areas in 2016). But they don't keep records about any private
> organisations (like shops, restaurants, ...), which are also very
> interesting for OSM.
>> Best regards
>> Michael
> Regards,
> Sander
> _______________________________________________
> Imports mailing list
> Imports at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/imports
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20151022/5971d9e3/attachment-0001.html>

More information about the Imports mailing list