Regarding this example:
https://github.com/gboeing/osmnx-examples/blob/master/notebooks/13-isolines-isochrones.ipynb
I propose a new method, make_iso_polys, outlined below. This, rather than a convex hull, might best demonstrate coverage area while preserved nuances ("inlets" or areas of inaccessibility not captured when creating a convex hull around a point cloud).
Example of what improved plot looks like:
Ful function defintion.
def make_iso_polys(G, edge_buff=25, node_buff=50):
isochrone_polys = []
for trip_time in sorted(trip_times, reverse=True):
subgraph = nx.ego_graph(G, center_node, radius=trip_time, distance='time')
node_points = [Point((data['x'], data['y'])) for node, data in subgraph.nodes(data=True)]
nodes_gdf = gpd.GeoDataFrame({'id': subgraph.nodes()}, geometry=node_points)
nodes_gdf = nodes_gdf.set_index('id')
edge_lines = []
for n_fr, n_to in subgraph.edges():
f = nodes_gdf.loc[n_fr].geometry
t = nodes_gdf.loc[n_to].geometry
edge_lines.append(LineString([f,t]))
n = nodes_gdf.buffer(node_buff).geometry
e = gpd.GeoSeries(edge_lines).buffer(edge_buff).geometry
all_gs = list(n) + list(e)
new_iso = gpd.GeoSeries(all_gs).unary_union
isochrone_polys.append(new_iso)
return isochrone_polys
Further elaboration available here: http://kuanbutts.com/2017/12/16/osmnx-isochrones/
Regarding this example:
https://github.com/gboeing/osmnx-examples/blob/master/notebooks/13-isolines-isochrones.ipynb
I propose a new method,
make_iso_polys, outlined below. This, rather than a convex hull, might best demonstrate coverage area while preserved nuances ("inlets" or areas of inaccessibility not captured when creating a convex hull around a point cloud).Example of what improved plot looks like:
Ful function defintion.
Further elaboration available here: http://kuanbutts.com/2017/12/16/osmnx-isochrones/