I am trying to extract pedestrian crossing data for Australia. When I look at OSM Cycle Map I can see the yellow points that seem to be exactly what I am after. However, there are a bunch of additional points in my Geofabrik dataset for fclass = crossing, code 5204. When I click query features on the OSM map in the approximate area of these additional points, they do come up as crossings (even though not visualised on the Cycle Map). Is there a hidden subclass somehow or how does this work? How can Cycle Map filter these out when they are all the same fclass and code?
I suppose that either the crossing=unmarked or the crossing:markings=no on the first makes it loose the yellow dot on the Cycle Map layer.
It’s not the crossing=traffic_signals on the other two, since a crossing I know well close to my home has a yellow dot and is tagged with crossing=uncontrolled + crossing:island=no + highway=crossing + kerb=lowered+tactile_paving=no`
I have no idea about the “fclass=crossing, code 5204” in the Geofabrik extracts, though…
Geofabrik shapefiles do not contain the full detail of OpenStreetMap (it would not be possible to represent OSM’s arbitrary number of tags in a shape file). Instead, the shapefiles only evaluate the most frequently used tags and put them in a well-documented structure.
If you are after the full detail of OSM, you will need to process the original OSM data and not the shape files. Get the .osm.pbf file, import it into a PostGIS database with osm2pgsql (be sure to use the --hstore option), and you will have access to the full breadth of OSM data.
The code 5204 the Geofabrik shapefiles is applied to anything that has a highway=crossing or railway=level_crossing tag in OSM.
