Skip to content

Network Analysis Methodology

Introduction

The core logic for the accessibility analysis is encapsulated within the RoutableNetwork class.

This code builds upon the pandana library, adding the ability to read data inputs from PostgreSQL and to save analysis results back to PostgreSQL.

Workflow

Step 1: Setup

  • Get a list of all unique POI categories
  • Assign start/end node ID values to each segment
  • Calculate the travel time cost to traverse each segment

Step 2: Construct network

  • Extract edge and node geodataframes from SQL
  • Build a pandana.Network() and precompute it

Step 3: Analyze POIs

  • Calculate each POI category by adding the POI category to the network and computing distances to the N-nearest POIs for every node in the network
  • Combine the results of each single POI analysis into one table
  • Save table to database and generate geotable with results tied to geometries
  • Clean up by moving results to a new schema and deleting temp QAQC tables

Python usage

To use this code directly in a Python process, import the RoutableNetwork class from the network_routing module, instantiate the class with keyword arguments, and run the command to compute every POI result into a single output table in postgres:

from network_routing import pg_db_connection
from network_routing.accessibility.routable_network import RoutableNetwork

db = pg_db_connection()

arguments = {
    "edge_table_name": "pedestriannetwork_lines",
    "node_table_name": "nodes_for_sidewalks",
    "node_id_column": "sw_node_id",
    "poi_table_name": "regional_transit_with_accessscore",
    "poi_id_column": "category",
    "output_table_name": "regional_transit_stops",
    "output_schema": "sw_defaults",
    "max_minutes": 120,
    "poi_match_threshold": 152,
}

net = RoutableNetwork(db, **arguments)
net.compute_every_poi_into_one_postgres_table()