Some person on reddit asked how they could use QGIS to calculate the number of police stations within 5 miles of a school. I made them a small tutorial.

draw_shortest_path_from_origin_to_destination.py

from collections import defaultdict

# Initialize layer variables
streets_lyr = QgsProject.instance().mapLayersByName('streets')[0]
origin_lyr = QgsProject.instance().mapLayersByName('schools')[0]
destination_lyr = QgsProject.instance().mapLayersByName('police_stations')[0]
od_matrix = QgsProject.instance().mapLayersByName('od_pairs')[0]

# Initialize variables for analysis
origin_ids = QgsExpression('array_agg("fid")').evaluate(origin_lyr.createExpressionContext())
destination_id_closest_origins = defaultdict(list)

# Make dict of lists of origin-destination pairs grouped by destination
for origin_id in origin_ids:
    closest_destinations = QgsExpression(
        f'array_agg(\
            "destination_id",\
            filter:="origin_id" = {origin_id},\
            order_by:="network_miles"\
        )').evaluate(od_matrix.createExpressionContext())
    destination_id_closest_origins[closest_destinations[0]].append(origin_id)

# Run "Shortest Path (Layer to Point)" processing on each o-d pair list
# See: https://docs.qgis.org/testing/en/docs/user_manual/processing_algs/qgis/networkanalysis.html#shortest-path-layer-to-point
all_paths = []
for destination_id in destination_id_closest_origins:
    origin_lyr.selectByIds(destination_id_closest_origins[destination_id])
    print(f'processing destination: {destination_id} from origins: {origin_lyr.selectedFeatureIds()}...')
    processing.run("native:shortestpathlayertopoint", {
        'DEFAULT_DIRECTION' : 2, 
        'DEFAULT_SPEED' : 50, 
        'DIRECTION_FIELD' : None, 
        'END_POINT' : destination_lyr.getFeature(destination_id).geometry().asPoint().toString(),
        'INPUT' : streets_lyr.id(),
        'OUTPUT' : f'{destination_id}',
        'SPEED_FIELD' : None,
        'START_POINTS' : QgsProcessingFeatureSourceDefinition(origin_lyr.id(), True),
        'STRATEGY' : 0,
        'TOLERANCE' : 0,
        'VALUE_BACKWARD' : '',
        'VALUE_BOTH' : '',
        'VALUE_FORWARD' : ''})
    all_paths.append(f'{destination_id}.gpkg')

# Load results as temporary layer
processing.runAndLoadResults("native:mergevectorlayers", {
    'LAYERS':all_paths,
    'CRS': iface.crs(),
    'OUTPUT':'memory'})

network_origin_destination_pairs_project.md

Data from OSM. police_stations = "amenity=police". schools = "amenity=school" excluding private religion schools and excluding preschools.

Table after origin-destination pairs analysis and processing.

Results visualized.

Results of draw_shortest_path_from_origin_to_destination.py

Full imgur album: https://imgur.com/a/6Y1LeX2