What I need to do is to select some of the petrol stations in The Netherlands, based on land use, proximity to cities and airports, wind speed, to name a few. Obviously, more accurate and updated data will give me a more realistic scenario.
I have a few questions about the OSM shapefiles:
1- How accurate are the POI, like petrol stations? I already downloaded some other POI databases for petrol stations (poi.gps-data-team.com/netherlands/petrol/) but I even had to fix some of the coordinates because a few stations were in the middle of the ocean. The total amount differs by hundreds, but I know some POI in the gps-data-team are repeated.
2- Is there another way to get the areas of all the airports besides the commercial version? Already spent a fair bit of time in http://www.nationaalgeoregister.nl/geonetwork/srv/dut/catalog.search, with the help of Google translate and had no luck. I got some of them using the Corine Land Cover, but if possible I’d like a more accurate shapefile.
Nice to hear you are considering to use OSM for your thesis. I did the same a while ago, although I just needed roads.
1: Now I’m not sure on how accurate the data for petrol stations is in the Netherlands. Looking at the spread of the fuel stations I see no real irregularities (it follow the population density more or less), but there are most likely missing ones.
2: Geofabrik surely is not the only way to get your data. Seems weird that they ask money for that data. You could do a simple overpass turbo query for aeroway=aerodrome and export the data as a GeoJSON. It’s a mixed bag of POIs and areas though and it does include small sports airports for gliders (indicated by zweef in the name), but it’s a fairly small dataset so manual filtering shouldn’t be too hard.
3: I’m not familiar with that site, but I only saw “bestemmingsplannen” a.k.a. Land-use planning maps. It might be that this is suitable for your research, it’d probably depend on its purpose. What kind of category are you interested in?
http://ourairports.com/data/ provides a csv with all airports of the world, including glider airfields, heliports and disused airports. At least it has a coordinate list, but no areas.
I myself use the raw osm country files from geofabrik and filter them using qgis. Qgis is free and has a lot of tools to export to shape, then you have the control.
About the POI:
I haven’t seem them for some time, but there used to be POI data from the AND import with amenity=fuel as node in a highway.
This was old data and I have removed a few in the past where I was sure there was no fuel station (anymore).
At other times, the real position was close by and the node on the highway was also removed.
Sometimes POI came from databases which mentioned addresses, not coordinates (e.g. amenity=post_box, supermarkets). These are not very accurate. But it looks to me things are improving quickly.
I agree with JanWandelaar in respect to POI’s and the AND import. I made a maproulette task for this a while ago, no idea what the status of that is at the moment.
I would suggest that when you have a POI with source=AND that you treat it with caution. For other POI’s you will have to accept that they are correct (the alternative is that you can not trust OSM data at all).
Thank you all from you messages. I’m using QGIS to work with all the data.
I think for now I’ll use both POI for fuel stations, and after I’ll filter them (based on proximity to different places, land use, etc). I’ll probably get a small number, making it easier to verify all of them with Google maps.