Getting started with OS NGD Transport Path features and pgRouting

The OS National Geographic Database (OS NGD) Transport Theme provides a definitive network dataset and topographic depiction of Great Britain’s roads, railways, tracks and paths. These notes provide an overview of how to create a routable network using the road and path featuers from the NGD Transport Theme with pgRouting. They are only intended to provide guidance and may need to be modified to suit specific use cases and system requirements.

PgRouting is an open-source tool that extends the spatial capabilities of the PostGIS extension to the PostgreSQL database to provide routing functions.

Software Requirements

PostgreSQL with the PostGIS and pgRouting extensions (pgRouting is packaged with PostGIS from version 2.1)

Data Requirements

• OS NGD Transport Network Road Link – network line geometry and attribution describing the link.

• OS NGD Transport Network Path Link – network line geometry and attribution describing the link.

• OS NGD Transport Network Connecting Link – line segmentation enabling the connection between road and path; it does not represent a real-world feature.

Before beginning you should also consider how much data you need, the bigger the area then the longer it will take to build and perform analysis.

Loading the data

The first step is to download the data from OS DataHub by using OS Select+Build to create a recipe. These are the table names used for this example, you will need to ensure they match or are substituted correctly when replicating.

OS NGD Collection and Feature Type	GeoPackage	Table name	pgRouting
OS NGD Transport Network Road Link	trn_ntwk_roadlink	trn_ntwk_roadlink	pgrouting.ngd_roadlink
OS NGD Transport Network Path Link	trn_ntwk_pathlink	trn_ntwk_pathlink	pgrouting.pathlink
OS NGD Transport Network Connecting Link	trn_ntwk_connectinglink	trn_ntwk_connectinglink	pgrouting.connectinglink

All OS NGD is available in the GeoPackage format which will be needed to load the data into our PostgreSQL database in order to be prepared for pgRouting. In this example we have created a database schema named ‘pgrouting’ and loaded the data using the DB Manager plugin in QGIS 3.8.1. When loading the data you will need to check the option to Create spatial index as these notes assume the data is indexed.

Preparing the data

Before building the network the 3 sets of links need to be combined into 1 layer, ensuring that they all have the same attribution. To keep this simple we have used the minimum attributes required from each. However, these need to be added to the connecting link as described in the following table:

endnode	varchar
startnode	varchar
endgradeseparation	Int4	0
startgradeseparation	Int4	0

These layers need to be populated, setting the grade separation as 0 and the connecting node and path node values can be used as the end and start node.-- grade separation is set to 0

update schema.connectinglink set endgradeseparation = 0;
update schema.connectinglink set startgradeseparation = 0;
-- start node to equal pathnode and end connecting node
update schema.connectinglink set endnode = connectingnodeid;
update schema.connectinglink set startnode = pathnodeid;

Once the attribution is populated the 3 layers can be merged into a new layer.

CREATE TABLE schema.roads_paths AS
SELECT a.geom, a.TOID, a.endnode, a.startnode, a.endgradeseparation, a.startgradeseparation, a.description
FROM  schema.ngd_roadlink  a
UNION ALL
SELECT b.geom, b.TOID, b.endnode, b.startnode, b.endgradeseparation, b.startgradeseparation, b.description
FROM schema.pathlink  b
UNION ALL
SELECT c.geom, c.TOID,  c.endnode, c.startnode, c.endgradeseparation, c.startgradeseparation, c.description
FROM schema.connectinglink c;

Add a geometry index to the table

--Geometry index
CREATE INDEX roads_paths_geom_idx
  ON schema.roads_paths 
  USING gist
  (geom);

Building the Network

These notes are for guidance, table and field names should be checked to ensure they match.

The first statement generates a node lookup using the startNode and endNode references.

Where appropriate, it introduces additional nodes using the grade separation to ensure the real-world physical separations between Links are handled accordingly.

CREATE TABLE schema.node_table AS
  SELECT row_number() OVER (ORDER BY foo.p) AS id,
         foo.p AS node
  FROM (
    SELECT DISTINCT CONCAT(a.startnode, a.startgradeseparation) AS p FROM schema.road_paths a
    UNION
    SELECT DISTINCT CONCAT(a.endnode, a.endgradeseparation) AS p FROM schema.road_paths a
  ) foo
  GROUP BY foo.p;

--Create an index on node table

CREATE INDEX node_table_node_idx
  ON schema.node_table
  (node);

The second statement creates the directed graph which can be used for routing.

It utilises the node lookup to generate the required ‘source’ and ‘target’ integer fields; along with adding 'cost' and 'reverse_cost', these re simply based on length and futher consideration will be needed depending on the types of road in your area. When building a network for vehicle routing you would incorporate one-ways at this point, however as we are creating a walking network these can be ignored.

It also uses the TOID field in the data to create an id for each link, however, to use this as an integer we have stripped OSGB from the start of the value.

CREATE TABLE schema.edge_table AS
  SELECT row_number() OVER (ORDER BY ltrim(a.toid, 'osgb')) AS id, 
         a.description,
         b.id AS source,
         c.id AS target,
         CASE 
         	WHEN description = 'Connecting Link' THEN st_length(a.geom)
         	WHEN description LIKE 'Path%' THEN st_length(a.geom)
         	WHEN EXISTS (SELECT r.presenceofpavement_overall_m FROM pgrouting.ngd_roadlink r 
         	WHERE r.presenceofpavement_overall_m  > 0 )
         	THEN st_length(a.geom)
         	ELSE '-9999'
         END AS cost,
        CASE 
         	WHEN description = 'Connecting Link' THEN st_length(a.geom)
         	WHEN description LIKE 'Path%' THEN st_length(a.geom)
         	WHEN EXISTS (SELECT r.presenceofpavement_overall_m FROM   pgrouting.ngd_roadlink r 
         	WHERE r.presenceofpavement_overall_m  > 0 )
         	THEN st_length(a.geom)
         	ELSE '-9999'
         END AS reverse_cost,
         a.geom AS the_geom
  FROM pgrouting.roads_paths AS a
    JOIN pgrouting.node_table AS b ON CONCAT(a.startnode, a.startgradeseparation) = b.node
    JOIN pgrouting.node_table AS c ON CONCAT(a.endnode, a.endgradeseparation) = c.node;
--Create indexes
CREATE UNIQUE INDEX edge_table_id_idx ON schema.edge_table (id);
CREATE INDEX edge_table_source_idx ON schema.edge_table (source);
CREATE INDEX edge_table_target_idx ON schema.edge_table (target);

--Geometry index
CREATE INDEX edge_table_geom_idx
  ON schema.edge_table
  USING gist
  (the_geom);

This has created a network data set comprising the path data with roads that may have a footpath on either side, using the presence of pavement attribute which is part of the NGD Transport Schema 2 data.

Example usage of pgRouting

This query calculates how far you can walk in 5 minutes and could be used to identify what facilities are available within walking distance of an address. The first step is to use the network we have created to calculate the walking distance.

CREATE TABLE schema.walk_15mins as
SELECT * FROM pgr_drivingDistance(
    'SELECT id, source, target, cost, reverse_cost FROM schema.edge_table',
    15879,  --source node id from the edge table
    900, --max time in seconds
    true --directed is true
);
--join to the edge table
CREATE TABLE schema.walk_15min_edges as
SELECT * FROM schema.edge_table as a 
join schema.walk_15mins as b on a.id = b.edge;

Create a Polygon Isochrone for your results

The results of this query are presented as a list of roads and paths that are reachable in the time specified. However, you may want to display isochrones for your output and for use in service area analysis. This query takes the output and creates a polygon that represents its boundaries using a concave hull.

CREATE TABLE schema.isochrone as
SELECT St_ConcaveHUll(St_collect(the_geom), 0.5) as geom
from schema.walk_15min_edges;

Points within the isochrone

Finally, to identify what is within the 15 minutes an intersect query can be used. In this example we are identifying all Community buildings near-by.

CREATE TABLE schema.community
SELECT * FROM schema.addressbaseplus AS a, schema.isochrone AS b
WHERE ST_intersects(a.geom, b.geom);

The results can then be displayed graphically in a GIS.

isochrone map created from pgRouting

Further information

https://github.com/mixedbredie/highways-for-pgrouting

https://github.com/tmnnrs/highways-network-pgrouting

PostGreSQL - https://www.postgresql.org/

PostGIS - http://www.postgis.net/

pgRouting - https://pgrouting.org/

http://planet.qgis.org/planet/tag/pgrouting/