General

This repository contains the scripts (written in Python 3) used for producing the results for the manuscript: "Travel times and transfers in public transport networks: comprehensive accessibility analysis based on Pareto-optimal journeys"

License

All scripts (under scripts/) are provided under the MIT license. (See scripts/LICENSE).

Acknowledgment

If you use these scripts, please cite our paper or give appropriate acknowledgement otherwise. (Details on citing the paper will be added once the paper has been published.)

Data

The data for running the analyses are provided under data/ The original source data (events and stops) for our analyses consists of GTFS-data and OpenStreetMap extracts:

• The original GTFS data has been downloaded from the Reittiopas API, and is provided under data/raw/
• The OpenStreetMap data was downloaded from Geofabrik's download server (Open Database License 1.0., not provided in the repository due to large file size).

Dependencies

The preprocessing of the source data, and all scripts in this repository use the gtfspy Python package (https://github.com/CxAalto/gtfspy), which is also included as an explicit git-submodule dependency in this repository. Note that, since the creation/run of the analysis scripts in these repository, the gtfspy package has most likely improved / changed, and thus this repository should not be considered as a long-term model example how to work with gtfspy. For introduction on how to use gtfspy, see the documentation in that package.

Moreover, the code in this repository is not necessarily compatible with the most recent version of the gtfspy package. The only version of gtfspy package that these scripts are guaranteed to work with is the accompanying git-submodule dependency.

Other (more standard) dependencies include:

• smopy (https://github.com/rossant/smopy)
• pandas
• numpy
• matplotlib
• Cython
• requests

Scripts

Modules used by all / most analysis scripts

• prepare.py
  • Add extra stops (or "locations" ) to the schedule database + create updated extracts.

Modules used by all / most analysis scripts

• settings.py
  • Shared settings for the analyses
• compute.py
  • Shared computation and caching pipelines for the analyses.
• util.py
  • Miscellanous shared utility functions.

Analyzes

• plot_one_day_example_profile.py
  • Compute and plot one long (6AM-9PM) temporal distance profile augmented with boarding-counts.
• plot_temporal_distance_profiles.py
  • Compute and plot two examples of temporal distance profiles with and without boarding-count information.
• plot_transfers_on_map.py
  • Compute and plot boardings counts on a map.
• temporal_distances_map.py
  • Compute and plot temporal distance statistics on a map.
• multiple_targets_fig.py
  • Compute and plot mean temporal distance and mean number of transfers on fastest-paths towards multiple locations.

All-to-all analyses

• all_to_all_slurm.sh
  • A script for submitting a batch job to Triton cluster (at Aalto University) for performing all-to-all analyses (calls compute_all_to_all_stats.py)
• compute_all_to_all_stats.py
  • A short script steering the all-to-all computations in Python
• analyze_all_to_all_stats.py
  • Analyze the results produced by compute_all_to_all_stats.py
• slurm_submit_command.txt
  • A reminder how to submit the batch job to Triton

Other analyses

• plot_temporal_distance_profiles_simple_verification.py
  • Plot fastest-path temporal distance profiles using a simple profiler that only tracks time.
• plot_profiles_on_a_map.py
  • Mostly old code that is no longer used.

Other plots

• schematic_plots.py
  • Plot schematic examples of temporal distance profiles.
• schematic_temporal_network_image.py
  • Plot a crude node-time diagram of a list of events.
• plot_route_map.py
  • Plot a route map showing the PT organization.

To rerun the analyses

0. python scripts/prepare.py
1. python scripts/plot_route_map.py
2. Run the scripts found listed above under Scripts -> Analyses
3. Run the scripts found listed above under Scripts -> All-to-all Analyses

Author / Point of contact:

Rainer Kujala ([email protected]; @rmkujala)