Shortest Path Routing with PostGIS - PGRouting - Network Analysis - Complete Web Project - Part 2

In this second part of the series, we focus on creating the node layer — a crucial component of any routing system.

Using the clean road line data (grid_lines) from Part 1, we extract key intersection points and line vertices, which will act as nodes in our network graph. These nodes are the building blocks for connecting paths and calculating routes.

✅ What’s Covered in this Video:

• Create a network_nodes table
  A dedicated table is created to store all points (nodes) that represent intersections or line segment vertices.

• Extract vertices from each LineString
  With ST_DumpPoints(), we extract every start/end and bend point of each line segment, converting them into point geometries.

• Detect & insert intersection nodes
  Additional nodes are generated from actual road intersections using ST_Intersection() — critical for building a real-world network.

• Deduplicate nodes
  Since many lines may share points, we clean the node list using DISTINCT ON (geom) to ensure each node appears only once.

• Apply spatial index
  A GIST index is created on the network_nodes table for fast nearest-neighbor and routing operations.

💡 Why This Step Matters:

Every pathfinding algorithm — like Dijkstra — works on a graph made of nodes and edges. This step gives life to the "nodes" of our graph, accurately reflecting the geometry and connectivity of the real-world road network.

In the next part, we'll use these nodes to build edges (connections) between them and assign costs based on real-world distances.