Package OSMnx should be investigated, as it proposes a very convenient gdf_from_place function.
(Although it fetches only one polygon, the source could serve as a good example for similar functionality.
Alternatively, as the NetworkX integration has been already started by #22, the whole shapefile phase could be skipped, with its ox.graph_from_place('Manhattan, New York, USA', network_type='drive') function. And the spatial data (at least vertex and edge, without demand and source data) could be calculated from that source.
There are even some possibilities to simplify the resulting graph, which could be an alternative to the current skeletron based approach.
Furthermore, this package already solved the automated UTM code calculation so at least, that would be of great use for us. :
>>> osmnx as ox
>>> city = ox.gdf_from_place('Berkeley, CA')
>>> city_xy = os.project_gdf(city)
>>> G = ox.graph_from_place('Manhattan, New York, USA', network_type='drive'
Summary:
Package OSMnx should be investigated, as it proposes a very convenient
gdf_from_placefunction.(Although it fetches only one polygon, the source could serve as a good example for similar functionality.
Alternatively, as the NetworkX integration has been already started by #22, the whole shapefile phase could be skipped, with its ox.graph_from_place('Manhattan, New York, USA', network_type='drive') function. And the spatial data (at least vertex and edge, without demand and source data) could be calculated from that source.
There are even some possibilities to simplify the resulting graph, which could be an alternative to the current
skeletronbased approach.Furthermore, this package already solved the automated UTM code calculation so at least, that would be of great use for us. :
Summary:
gdf_from_place(), let it inspire us.OSM -> NetworkX -> GeoDataFrameworkflow instead of / besidesOSM -> Mapzen -> Shapefiles -> GeoDataFrame