Developing a cycling infrastructure classification system for Greater Melbourne, Australia using OpenStreetMap

Background

Policy-makers are looking to promote the uptake of cycling as a healthy mode of travel that reduces the negative effects of traditional motorised transport (physical inactivity, air pollution, traffic congestion) and achieves sustainability goals. As an active form of mobility, bicycling improves physical and mental health and has long-term public health benefits. However, there are a number of barriers that prevent people from riding a bike, including fears about riding alongside motor vehicle traffic and the lack of safe and appropriate bicycling infrastructure. For the strategic installation of safer cycling infrastructure or the improvement of existing infrastructure, rigorous evidence-informed scientific studies are necessary, which in turn rely on high-quality bicycling data, which is scarce.

In this regard, one of the prerequisites is understanding the different types of cycling infrastructure that exist in an urban area and create an inventory dataset that can form the basis of future cycling-related research. OpenStreetMap (OSM) is a valuable open-source map database that contains transport infrastructure data among other things and has spatial coverage for almost the entire planet. Hence, it is used extensively by researchers and planners and it helps develop methods that are transferable and thus can be replicated irrespective of the study area. We, the Sustainable Mobility and Safety Research Group (SMSR) at Monash University, Australia, have developed a classification process to classify existing bicycling infrastructure across Greater Melbourne, Australia. We have derived knowledge from existing studies and calibrated our classification system to suit local tagging practices.

Update

1. Roads which have associated bike lanes (usually protected) which are mapped separately by OSM contributors, should not be classified as mixed traffic (as was the case in the earlier version). Thus, we have decided to classify roads with associated bike infra mapped separately in a new category '0b. Associated bike infrastructure separately mapped'. However, it is not trivial to detect the associated bike lane/infrastructure type. Also, given that infrastructure is separately mapped and is classified already, it could cause duplication issues.
2. Some protected bike lanes (e.g. along St Kilda Road) were getting incorrectly classified as 'Dedicated bikepaths' given that they are mapped separately to the associated road in OSM. We can now classify such instances correctly.
3. We have also added a 'superclass' classification, which is a less granular classification, and clubs together similar but separate classes from the more granular classification system.

Technology

• Python
• OpenStreetMap
• Python packages
  • OSMnx
  • NetworkX
  • geojson
  • Shapely
  • Pandas

How to implement in Python

• Install the aforementioned Python packages as described in the links.
• A settings.py file has been provided. This file contains a more holistic set of OSM tags that are not retrieved by default by OSMnx. Copy the contents of this file into your settings.py file under your osmnx folder/installation. Otherwise, the code will throw KeyError.
• Run the Python codes provided.

Sample results

Distribution of different types of bicycling infrastructure at the link-level in Greater Melbourne.

Distribution of different types of bicycling infrastructure at the link-level in Inner Melbourne.

Distribution of bike paths and protected bike lanes in Inner Melbourne.

Authors

The document has been prepared by Dr Debjit Bhowmick and Assoc Prof Ben Beck from the Sustainable Mobility and Safety Research (SMSR) Group at Monash University and Paul Yacoumis from the Victorian Department of Transport and Planning. For any queries, please contact Dr Debjit Bhowmick (Research Fellow, debjit.bhowmick@monash.edu) or Assoc Prof Ben Beck (Head of SMSR, ben.beck@monash.edu).

Citation

If you are using this classification system for your work, we strongly recommend citing this repository. You may refer to the guidelines provided here.

BibTeX

@misc{Sustainable_Mobility_and_Safety_Research_Group_Bicycling_infrastructure_classification_2023, author = {{Sustainable Mobility and Safety Research Group, Monash University}}, title = {{Bicycling infrastructure classification using OpenStreetMap}}, url = {https://github.com/SustainableMobility/bicycling-infrastructure-classification}, doi = {https://doi.org/10.6084/m9.figshare.27059983.v1}, year = {2023} }

APA

Sustainable Mobility and Safety Research Group, Monash University. (2023). Bicycling infrastructure classification using OpenStreetMap. https://github.com/SustainableMobility/bicycling-infrastructure-classification

Note: OSM is a volunteered geographic information and is prone to occasional completeness and correctness issues, especially in the case of bicycling infrastructure due to inconsistent tagging practices. This can lead to occasional misclassification, especially if directly translated to other study areas, especially outside Australia.