Skip to Content
You are currently on the new version of our website. Access the old version .
Urban ScienceUrban Science
  • Editor’s Choice
  • Article
  • Open Access

10 August 2020

20 Pages

Spatial Factor—Using a Random Forest Classification Model to Measure an Internationally Comparable Urbanity Index

,
,
,
,
and
1
BMW AG, Petuelring 130, 80788 Munich, Germany
2
Institute for Transport Studies, Karlsruhe Institute of Technology (KIT), Kaiserstrasse 12, 76131 Karlsruhe, Germany
*
Authors to whom correspondence should be addressed.

Abstract

Travel behavior can be determined by its spatial context. If there are many shops and restaurants in close proximity, various activities can be done by walking or cycling, and a car is not needed. It is also more difficult (e.g., parking space, traffic jams) to use a car in high-density areas. Overall, travel behavior and dependencies on travel behavior are influenced by urbanity. These relationships have so far only been examined very selectively (e.g., at city level) and not in international comparison. In this study we define an Urbanity Index (UI) at zip code level, which considers factors influencing mobility, international comparability, reproducibility as well as practical application and the development of a scalable methodology. In order to describe urbanity, data were collected regarding spatial structure, population, land use, and public transport. We developed the UI using a supervised machine learning technique which divides zip codes into four area types: (1) super-urban, (2) urban, (3) suburban/small town, (4) rural. To train the model, the perception from experts in known zip codes concerning urbanity and mobility was set as ground truth. With the UI, it is possible to compare countries (Germany and France) with a uniform definition and comparable datasets.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.