routing

ADVANCED BETA

This module contains functions that perform routing and path calculations natively in BigQuery without the need of calling external location data services. In order to run the functions of this module the user needs to have access to CARTO's pre-generated road network (based on OSM segments) that is available as a public subscription via the Data Observatory. Please check this guide to learn how to subscribe to a dataset from the Data Observatory.

ROUTING_MATRIX

ROUTING_MATRIX(start_point_array, dest_point_array, area_of_interest, transportation_mode, do_network_table, do_source, output_table, options)

Description

This procedure calculates the shortest paths in terms of travel times or distances for all routes between all of a given set of locations. It requires a Data Observatory road network subscription to perform the calculations.

For every given origin, this procedure calculates the minimum cost of travel from that origin to every given destination on the road network specified.

  • start_point_array: ARRAY<GEOGRAPHY> Source points array. the node of the network nearest to this point will be used as the source point to compute the shortest path.

  • dest_point_array: ARRAY<GEOGRAPHY> destination points array. the node of the network nearest to this point will be used as the destination point to compute the shortest path.

  • area_of_interest: GEOGRAPHY area of interest over where the analysis takes place.

  • transportation_mode: STRING type of transportation mode to be used for the calculation of routes. Available options: car, car_motorway_only, car_major_road_only, bicycle or foot.

  • do_network_table: STRING identifier (slug) of the Data Observatory Network table.

  • do_source: STRING name of the location where the Data Observatory subscriptions of the user are stored, in <my-dataobs-project>.<my-dataobs-dataset> format. If only the <my-dataobs-dataset> is included, it uses the project carto-data by default. It can be set to NULL or ''.

  • output_table: STRING the full path name of the output table.

  • options: STRING containing a valid JSON with the different options. Valid options are described the table below. If options is set to NULL the all options are set to default.

Return type

The output table includes the following columns:

  • start_geo: GEOGRAPHY Start point from source points array.

  • dest_geo: GEOGRAPHY Destination point from destination points array.

  • start_geo_snapped: GEOGRAPHY Start point snapped to the nearest start node of links of the network.

  • dest_geo_snapped: GEOGRAPHY Destination point snapped to the nearest destination node of links of the network.

  • start_order: INT64 Start point position in the source points array.

  • dest_order: INT64 Destination point position in the destination points array.

  • start_s2: INT64 Unique identifier of the start point snapped from start point.

  • dest_s2: INT64 Unique identifier of the destination node snapped from destination point.

  • cost: FLOAT64 Overall cost of the path (travel time or distance depending on the TYPE parameter value).

  • distance: FLOAT64 Overall driving distance of the path in meters.

  • travel_time: FLOAT64 Overall travel time of the path in seconds.

  • path: GEOGRAPHY Overall path.

  • detailed_linestring: RECORD: Array of links that makes up the path.

    • start_s2 INT64: Unique identifier of the start node of the link.

    • dest_s2 INT64: Unique identifier of the destination node of the link.

    • speed: FLOAT64 Speed over the link.

    • cost: FLOAT64 Cost of the link (travel time or distance depending on the TYPE parameter value).

    • distance: FLOAT64 Driving distance of the link in meters.

    • travel_time: FLOAT64 Travel time of the path in seconds.

    • path: GEOGRAPHY Path of the link.

    • detailed_geography: ARRAY<GEOGRAPHY> Array of points that makes up the link.

Example

CALL `carto-un`.carto.ROUTING_MATRIX(
     -- start_point_array
    [ST_GEOGPOINT(-73.0, 40.0),ST_GEOGPOINT(-73.0, 41.0)],
    -- dest_point_array
    [ST_GEOGPOINT(-75.0, 41.0),ST_GEOGPOINT(-75.0, 40.0)],
    -- area of interest,
    ST_GEOGFROMTEXT("FULLGLOBE"),
    -- transportation mode
    'car',
    -- do_network_table
    'cdb_road_networ_81badfc2',
     --do_source
    '<my-dataobs-project>.<my-dataobs-dataset>',
     --output_table
    '<my-project>.<my-dataset>.<output_filename>',
    -- options
    """
    {
       "TYPE":"time",
       "MAX_COST":"100000",
       "WITH_PATH":"True"
    }
    """
);
-- {
--   "start_geo": "POINT(-73 40)",
--   "dest_geo": "POINT(-75 41)",
--   "start_geo_snapped": "POINT(-74.013134 40.688339)",
--   "dest_geo_snapped": "POINT(-74.026365 40.685995)",
--   "start_order": "0",
--   "dest_order": "0",
--   "start_s2": "-8520148151882761037",
--   "dest_s2": "-8520148044108704841",
--   "cost": "1082.6350274098224",
--   "distance": "1397.2114843386769",
--   "travel_time": "1082.6350274098224",
--   "path": "LINESTRING(-74.013134 40.688339, ..., -74.026365 40.685995)",
--   "detailed_linestring": [{
--     "start_s2": "-8520148151882761037",
--     "dest_s2": "-8520148151714079533",
--     "speed": "1.9",
--     "cost": "10.694765667341704",
--     "distance": "20.320054767949237",
--     "travel_time": "10.694765667341704",
--     "path": "LINESTRING(-74.013134 40.688339, -74.013375 40.688339)",
--     "detailed_geography": ["POINT(-74.013134 40.688339)", "POINT(-74.013375 40.688339)"]
--   },
--   ...,
--   {
--     "start_s2": "-8520148151867093047",
--     "dest_s2": "-8520148151854368777",
--     "speed": "1.2",
--     "cost": "11.348771839149197",
--     "distance": "13.618526206979036",
--     "travel_time": "11.348771839149197",
--     "path": "LINESTRING(-74.013587 40.688492, -74.013702 40.688578)",
--     "detailed_geography": ["POINT(-74.013587 40.688492)", "POINT(-74.013702 40.688578)"]
--   }]
-- }, {
--   "start_geo": "POINT(-73 40)",
--   "dest_geo": "POINT(-75 40)",
--   "start_geo_snapped": "POINT(-74.013134 40.688339)",
--   "dest_geo_snapped": "POINT(-74.026041 40.684658)",
--   "start_order": "0",
--   "dest_order": "1",
--   "start_s2": "-8520148151882761037",
--   "dest_s2": "-8520148043684863917",
--   "cost": "1124.4760381433846",
--   "distance": "1248.9194302355879",
--   "travel_time": "1124.4760381433846",
--   "path": "LINESTRING(-74.013134 40.688339, ..., -74.026041 40.684658)",
--   "detailed_linestring": [{
--     "start_s2": "-8520148151882761037",
--     "dest_s2": "-8520148151721556507",
--     "speed": "1.5",
--     "cost": "19.218684425894505",
--     "distance": "28.828026638841756",
--     "travel_time": "19.218684425894505",
--     "path": "LINESTRING(-74.013134 40.688339, -74.013221 40.688225, -74.013358 40.688151)",
--     "detailed_geography": ["POINT(-74.013134 40.688339)", "POINT(-74.013221 40.688225)", "POINT(-74.013358 40.688151)"]
--   },
--   ...,
--   {
--     "start_s2": "-8520148055505489297",
--     "dest_s2": "-8520148043684863917",
--     "speed": "1.0",
--     "cost": "305.40490363182045",
--     "distance": "305.40490363182045",
--     "travel_time": "305.40490363182045",
--     "path": "LINESTRING(-74.022772 40.684401, ..., -74.026041 40.684658)",
--     "detailed_geography": ["POINT(-74.022772 40.684401)", ..., "POINT(-74.026041 40.684658)"]
--   }]
-- },
-- ...

Limitations

Since this module runs natively on Bigquery, it relies exclusively on the resources allocated by the data warehouse for the query.

If a request fails due to a resource limit, you can try the following:

  • reduce the size of the network (reduce the size of the area of interest)

  • reduce or split into different queries the points in start_point_array

  • set or reduce (if it already exists) the MAX_COST parameter

  • set WITH_PATH parameter to False

In some cases road networks contain segments that are not connected to the main network. If any of the destination or origin points are closer to such segments than to other parts of the network it won't be possible to find routes to or from such points and some paths can be missing from the results. We're working on improving the quality of the road networks to avoid such problems. If you find this problem using the car transportation mode you can try using car_major_road_only or car_motorway_only instead, since the major roads network is less prone to this kind of problem.

ROUTING_ISOLINES

ROUTING_ISOLINES(start_point_array, cost_limit_array, area_of_interest, transportation_mode, do_network_table, do_source, output_table, options)

Description

This procedure generates a table containing isolines in terms of either travel times (isochrones) or distances (isodistances) for a given set of origin locations and range limits to be computed on the road network table.

  • start_point_array: ARRAY<GEOGRAPHY> Source points array. The node of the network nearest to this point will be used as the source point to compute the shortest path.

  • cost_limit_array: ARRAY<FLOAT64> Cost limit array. For each cost limit all the path within this range are returned. For an isochrone the cost is time in seconds, for an isodistance it's distance in meters.

  • area_of_interest: GEOGRAPHY Area of interest over where the analysis takes place.

  • transportation_mode: STRING Type of transportation mode to be used for the calculation of isolines. Available options: car, car_motorway_only, car_major_road_only, bicycle or foot.

  • do_network_table: STRING Identifier (slug) of the Data Observatory Network table.

  • do_source: STRING Name of the location where the Data Observatory subscriptions of the user are stored, in <my-dataobs-project>.<my-dataobs-dataset> format. If only the <my-dataobs-dataset> is included, it uses the project carto-data by default. It can be set to NULL or ''.

  • output_table: STRING The full path name of the output table.

  • options: STRING Containing a valid JSON with the different options. Valid options are described the table below. If options is set to NULL the all options are set to default.

Return type

The output table includes the following columns:

  • cost_limit: FLOAT64 Cost limit from the start_point_array input parameter taken in account for this row.

  • cost_limit_idx: INT64 Cost limit position from the start_point_array input parameter taken in account for this row.

  • start_order: INT64 Start point position from the source points array taken in account for this row.

  • start_geo: GEOGRAPHY The point geometry of the starting node taken in account for this row.

  • start_geo_snapped: GEOGRAPHY Start point snapped to the nearest start node of links of the network.

  • start_s2: INT64 Index of the node snapped from start point.

  • start_cost: FLOAT64 Cost from snapped start point to the start node of the link.

  • dest_s2: INT64 The unique identifier of the destination node of the link.

  • dest_cost: FLOAT64 Cost from snapped start point to the destination node of the link.

  • detailed_geography: ARRAY<GEOGRAPHY> Array of points that makes up the link.

  • detailed_geography_chunked_agg: GEOGRAPHY Link geography chunked to cost limit and aggregated in a single geography.

Example

CALL `carto-un`.carto.ROUTING_ISOLINES(
   -- start_point_array
    [ST_GEOGPOINT(-73.0, 40.0)],
    -- cost_limit_array
    [60., 120.],
    -- area of interest,
    ST_GEOGFROMTEXT("FULLGLOBE"),
    -- transportation mode
    'car',
    -- do_network_table
    'cdb_road_networ_81badfc2',
     --do_source
    '<my-dataobs-project>.<my-dataobs-dataset>',
    --output_table
    '<my-project>.<my-dataset>.<output_filename>',
   -- options
   """
   {
      "TYPE":"time",
      "MAX_COST":"100000",
      "WITH_PATH":"False"
   }
   """
);

-- {
--   "start_geo": "POINT(-122.3296557 47.582691)",
--   "dest_geo": "POINT(-122.3290612 47.5807722)",
--   "start_geo_snapped": "POINT(-122.3296557 47.582691)",
--   "dest_geo_snapped": "POINT(-122.3290612 47.5807722)",
--   "start_order": "0",
--   "dest_order": "35164",
--   "start_s2": "6093487519859207645",
--   "dest_s2": "6093440836742041863",
--   "cost": "27.095695089725034",
--   "distance": "309.31565973502575",
--   "travel_time": "27.095695089725034",
--   "path": "LINESTRING(-122.3296557 47.582691, -122.3298669 47.5827281, ..., -122.3290612 47.5807722)",
--   "detailed_linestring": [{
--     "start_s2": "6093487519859207645",
--     "dest_s2": "6093440836214032467",
--     "speed": "4.166666666666667",
--     "cost": "5.8301084779070846",
--     "distance": "24.292118657946187",
--     "travel_time": "5.8301084779070846",
--     "path": "LINESTRING(-122.3296557 47.582691, -122.3298669 47.5827281, -122.3299725 47.5827299)",
--     "detailed_geography": ["POINT(-122.3296557 47.582691)", "POINT(-122.3298669 47.5827281)", "POINT(-122.3299725 47.5827299)"]
--   },
--   ...,
--   {
--     "start_s2": "6093440836753434943",
--     "dest_s2": "6093440836742041863",
--     "speed": "18.055555555555554",
--     "cost": "0.658359285678671",
--     "distance": "11.887042658087113",
--     "travel_time": "0.658359285678671",
--     "path": "LINESTRING(-122.3290623 47.5808791, -122.3290612 47.5807722)",
--     "detailed_geography": ["POINT(-122.3290623 47.5808791)", "POINT(-122.3290612 47.5807722)"]
--   }]
-- }, ...

The output table returns links that form the isolines indexed by origin location and range limits. To get the full isoline geographies you need to aggregate the table:

SELECT ST_UNION_AGG(detailed_geography_chunked_agg) AS full_isoline_geography
FROM `output_table`
GROUP BY cost_limit, start_s2

Limitations

Since this module runs natively on Bigquery, it relies exclusively on the resources allocated by the data warehouse for the query.

If a request fails due to a resource limit, you can try the following:

  • reduce the size of the network (reduce the size of the area of interest)

  • reduce or split into different queries the points in start_point_array

  • reduce the maximum the cost limit in cost_limit_array

In some cases road networks contain segments that are not connected to the main network. If any of the destination or origin points are closer to such segments than to other parts of the network it won't be possible to find routes to or from such points and some paths can be missing from the results. We're working on improving the quality of the road networks to avoid such problems. If you find this problem using the car transportation mode you can try using car_major_road_only or car_motorway_only instead, since the major roads network is less prone to this kind of problem.

Last updated