Computation of service areas

jwestw commented 3 years ago

Create euclidean buffer around the transport stops
combine the polygons created
plot the polygons

james-westwood commented 3 years ago

The buffer function is working. Pushed in this commit https://github.com/james-westwood/SDG_11.2.1/commit/dd720361ca0b58f3858b9ed017c881cd5404c8e5

EDIT: now I am not sure this is working because although it is returning polygons which are bigger than the points, they do not seem to be 0.5km across.

james-westwood commented 3 years ago

Trying to get the buffered stops to join up. Have tried .dissolve which uses Shapley's unary_union method in the background.

So just tried it with unary_union directly. This is the output.

I think it would be better to niche down to a much smaller sub-section of B'ham and plot the stops (Points) and the service area (MULTIPOLYGON) to see the difference.

EDIT: with the buffer set to 500, this is the plot. Clearly this is not 500km otherwise there would be no gaps.

james-westwood commented 3 years ago

Although the dots on the plot above are polygons, I think they are much too small (though i don't know how to establish how big/small they are in km). I need help from Geospatial here. The polygons need to be 0.5km across.

james-westwood commented 3 years ago

Isolating a small sub section of Birmingham. Using a data set from https://www.doogal.co.uk/ I am isolating the stops in Acocks Green, a ward of Birmingham. I am doing this to make calculations easier.

# grab some coordinates for a little section of birmingham
ag_east_north = pd.read_csv('Acocks_Green_postcodes.csv', usecols=['Easting', 'Northing'])

eastings = [easting for easting in ag_east_north.Easting]
northings = [northing for northing in ag_east_north.Northing]
east_min = min(eastings)
east_max = max(eastings)
north_min = min(northings)
north_max = max(northings)

# limit the stops, filtering by the min/max eastings/northings for Acocks Green
north_mask = (north_min < birmingham_stops_geo_df['Northing']) & (birmingham_stops_geo_df['Northing'] < north_max)
east_mask = (east_min < birmingham_stops_geo_df['Easting']) & (birmingham_stops_geo_df['Easting'] < east_max)
# Filter the stops 
ag_stops_geo_df = birmingham_stops_geo_df[north_mask & east_mask]

And plotting within Birmingham it's like this:

I feel like this is successfully calculating and plotting the service areas. However the buffer function just needs to be fixed to buffer to exactly 0.5km.

jwestw commented 2 years ago

Closing this old issue as it was done long ago.

ONSdigital / SDG_11.2.1

Computation of service areas #21