PTM Validation

Two notebooks have been retained for the process of undertaking PTM validation:

  • 01_undertake_network_validation.ipynb – designed to validate route coding, especially focused on removing looping within bus routes.
  • 02_undertake_model_validation.ipynb – designed to validate the PTM assignment against observed sources.

Network Validation

01_undertake_network_validation.ipynb

This notebook can be run before the model has been imported into EMME and offers an opportunity to explore network coding issues, such as path errors or looping of services within the model.

This notebook performs a network validation for lines built using input notebooks. The process involves creating shortest paths within the notebook and comparing them to the input lengths.

Outputs from this notebook are saved in both the schema sqlite database and within a geopackage (to allow for spatial checking using a package such as QGIS).

Here is a summary of the key steps and operations performed:

  1. Imports and Setup - The notebook begins by importing necessary libraries such as geopandas, pandas, plotly, networkx, and others. It sets up paths and logging configurations.
  2. User Defined Variables - A schema name is defined for the validation process.
  3. Import Schema Data and Network - The notebook reads nodes, links, itineraries, and lines from the PTM database and logs the number of each imported.
  4. Create Modal Network Graphs & Build Transit Lines - It builds modal network graphs for different modes (e.g., buses) and constructs transit shapes using shortest paths. Coding errors are collected during this process.
  5. Compare Lengths - The notebook calculates the lengths of the built transit shapes and compares them with the GTFS lengths. It computes deviations and categorises the coding status into three levels: likely no coding issue, likely coding issue, and likely significant coding issue.
  6. Check the Number of Lines by Coding Status - It visualises the number of lines by their coding status using a bar chart.
  7. Check the Distribution of Line Length Deviations - A histogram is created to show the distribution of length deviations by coding status.
  8. Identify Significant Issues - The notebook identifies lines with significant coding issues and visualises them using a grouped bar chart comparing model lengths and GTFS lengths.
  9. Export for Validation - Finally, it exports the length comparisons and line issues to a database and a GeoPackage file for further validation.

The images below show examples from the Network Validation dashboard. The first image provides a high-level overview of how many transit lines may have potential coding issues.

Network Validation Overview

The second image shows for a number of transit lines that the model length significantly exceeds the GTFS length, potentially indicating that “looping” may be occuring for some routes.

Network Validation Example

Model Validation

This notebook is designed to validate public transport models using various datasets. The notebook has been designed to show all of the validation results (including charts and plotting) within the same notebook. Outputs can also be imported into Tableau.

It takes the following inputs:

  • validation.gpkg
  • screenlines.gpkg
  • Model extract data

The outputs generated by the notebook, which are imported into Tableau Hyper, include:

  • Passenger boardings by mode and time period
  • Bus line boardings
  • Station boardings
  • Screenline volumes
  • Screenline GEH (a statistical measure)

The notebook performs several key functions:

  1. Defines user variables - Such as SCHEMA_NAME, EXTRACTIONS, and PT_VALIDATION_DATASET_PATH.
  2. Sets up file paths and dashboards - Initialises paths and validation results.
  3. Processes observed data - Reads observed stop movements, route boardings, bus screenlines, corridor explorer links, and rail screenlines.
  4. Obtains validation data - For each extraction, it processes segment statistics, compares passenger boardings, bus line boardings, bus screenlines, station boardings, and rail screenlines.
  5. Generates plots - Creates and displays plots for passenger boardings, bus line patronage, station boardings, bus screenlines, and rail screenlines.
  6. Exports validation data - Exports the results to a SQL database.