proposal of POI layer about openmaptiles HOT 16 OPEN

The iD Editor has a lot of metadata , maybe we can reuse ..

• https://github.com/openstreetmap/iD/tree/master/data/presets/presets

examples:

• amenity=biergarten -> "icon": "beer" -> [ class = beer ; POINT + AREA ]
  https://github.com/openstreetmap/iD/blob/master/data/presets/presets/amenity/biergarten.json

• "icon": "religious-jewish" = "amenity": "place_of_worship" + "religion": "jewish"
  https://github.com/openstreetmap/iD/blob/master/data/presets/presets/amenity/place_of_worship/jewish.json

• Homeless Shelter === amenity=social_facility + social_facility=shelter + social_facility:for=homeless
  https://github.com/openstreetmap/iD/blob/master/data/presets/presets/amenity/social_facility/homeless_shelter.json

They have categories: https://github.com/openstreetmap/iD/tree/master/data/presets/categories
and depracated lists ( to replace old taggings ) https://github.com/openstreetmap/iD/blob/master/data/deprecated.json

from openmaptiles.

For me the preferred document format is .csv ( can be read and edit with Excel /Libreoffice )

Well we could build support for this in openmaptiles-tools..and then generate functions (this code already exists for generating).

from openmaptiles.

if you think my proposal is reasonable. I would like to assigning the OSM tags into these buckets .and of course i will enrich the classification and the rank. it will contain information about traffic and other important information.

Go for it!!

could you give me the link about the tags of the OSM POI tags? I don't know which is the exactly correct one.

http://wiki.openstreetmap.org/wiki/Points_of_interest

A good resource is the Geofabrik paper on how to structure OSM tags (Chapter 4.2):
https://www.geofabrik.de/data/geofabrik-osm-gis-standard-0.6.pdf

from openmaptiles.

There's also https://github.com/simonpoole/beautified-JOSM-preset .

But let me insist in this: Did anybody look at ? https://github.com/geometalab/osmaxx/blob/develop/docs/osmaxx_data_schema.md#poi_p and the data schema of our OSMaxx project? OSMaxx is focussing on basemaps and analysis in GIS. But with basemaps from OSM data the two projects share a goal, so they can benefit from each other.

from openmaptiles.

I would be ok with these additional categories (1st and 2nd class). Certainly makes it easier
to assign the same icon to same classes than now.

The big work however is assigning the OSM tags into these buckets (100+tags).
Would welcome such work. I don't have strong opinions for the POI layer right now - just important that there is a ranking and perhaps later on also make it possible to put important landmarks and transport POIs into the layer.

from openmaptiles.

I like the idea of storing all information ( business logic ) in the single metadata for the POI layer ..
because - we separate to logic with the codes .. and it is easy to understand for the users

For me the preferred document format is .csv ( can be read and edit with Excel /Libreoffice )

Tilezen using similar metadata in excel :

• https://github.com/tilezen/vector-datasource/tree/master/spreadsheets

from openmaptiles.

I like the idea of storing all information ( business logic ) in the single metadata for the POI layer ..
because - we separate to logic with the codes .. and it is easy to understand for the users

Well yes..but then on the other hand it can be just as confusing where it now comes from.
We had it in a YAML file before but I put it into SQL to make it very explicit where it is defined.

See osm2vectortiles/osm2vectortiles#470.

from openmaptiles.

I like - If we can move all information to a single .csv file , ( in the future )

• but this is not a quick task
• and need a lot of design time ( for a good architecture ) and integrating to openmaptiles-tools

but I add here my ideas :

probably we can add a rank and classification metadata to this document

• https://github.com/openmaptiles/openmaptiles/blob/master/layers/poi/class.sql
• and imposm3 mapping infos https://github.com/openmaptiles/openmaptiles/blob/master/layers/poi/mapping.yaml

And we can extend with maki icon names ..

• https://github.com/mapbox/maki
• https://github.com/osm2vectortiles/osm2vectortiles/blob/master/src/import-sql/classes/maki_label.yml

And ( later ) we can add ( Y/N ) columns about thematic layers or styles

• Hiking
• Bicycle
• Humanitarian
• City
• Tourism
• Transport
• Liberty styles or other styles ( https://github.com/openmaptiles/openmaptiles-gl-styles )
• ...

So we can create more slim POI layers ( removing the unwanted POI-s ) by automatically ...

from openmaptiles.

probably we can add a rank and classification metadata to this document

Having more detailed ranking in a CSV is fine.
One could also use ordered SQL types.

And we can extend with maki icon names ..

I don't want to have support for maki icons in OpenMapTiles since this is really Mapbox specific.
What one can always do is rename the open Maki icon set to fit the classes/types of the POIs when designing a style.

This way one can use the Maki icon set but does not use the potentially copyrighted (imho you cannot do that) mapping from tags to Maki icons

from openmaptiles.

@lukasmartinelli @ImreSamu

big work however is assigning the OSM tags into these buckets (100+tags).

if you think my proposal is reasonable. I would like to assigning the OSM tags into these buckets .and of course i will enrich the classification and the rank. it will contain information about traffic and other important information.

could you give me the link about the tags of the OSM POI tags? I don't know which is the exactly correct one.

from openmaptiles.

@bearnxx :

probably you can use the OSM Taginfo statistics
like:

• http://taginfo.openstreetmap.org/keys/amenity#values
• http://taginfo.openstreetmap.org/keys/shop#values
• http://taginfo.openstreetmap.org/keys/sport#values
• http://taginfo.openstreetmap.org/keys/leisure#values
• ...

Extreme example :)

The ice_cream can be tagged more ways : http://taginfo.openstreetmap.org/search?q=ice_cream#values

• 9 067 : cuisine=ice_cream
• 8 780 : amenity=ice_cream
• 1 743 : shop=ice_cream

But checking with the Overpass ( “cuisine=ice_cream in iceland” ) http://overpass-turbo.eu/s/kaS
most of them

• amenity=cafe + cuisine=ice_cream
• amenity=fast_food + cuisine=ice_cream
• amenity=restaurant + cuisine=ice_cream

So only this 2 tagging version we can use :

• amenity=ice_cream
• shop=ice_cream

important

You can download the taginfo data http://taginfo.openstreetmap.org/download ( Sqlite databases )

• and it is easy to export to .csv file and import this to your working excel ( as a new schema )
  • need a minimal sql knowledge

And please read my review ( #15 ) , it was based on Taginfo

if you finished the first 5-10% please show us , and we can review ...
( or you can use https://docs.google.com/spreadsheets/ and share a link .. )

from openmaptiles.

See also our work on structuring POIs and others in our OSMaxx project https://github.com/geometalab/osmaxx/blob/develop/docs/osmaxx_data_schema.md#poi_p

from openmaptiles.

Get it! @ImreSamu @lukasmartinelli @sfkeller

from openmaptiles.

The iD Editor has a lot of metadata , maybe we can reuse ..

Yup I like that one very much (given it is not close to maki of Mapbox - have not checked that).

from openmaptiles.

@lukasmartinelli

Yup I like that one very much (given it is not close to maki of Mapbox - have not checked that).

it is easy to check :)
https://github.com/openstreetmap/iD/search?utf8=%E2%9C%93&q=maki

from openmaptiles.

I‘m a little busy recently. I have added some new class in the POI schema. look at the excel

github don't support csv file for attachment. so I upload xlsx instead.
what class can we add to the file?

@lukasmartinelli @ImreSamu @klokan @nvkelso

POI_Classification.xlsx

from openmaptiles.