A Comparative Analysis of Urban Transport Using K-Means Clustering and Multi-Class Classification

Abstract

The transportation planning process requires a comprehensive study of the regions that need development. This study is an extension of the methodology of transportation planning. The authors use real-time data from Foursquare API to map out the number of transportation facilities and infrastructure available for each city. This study will shed light on areas that need the most development in terms of intra-neighbourhood and inter-neighbourhood transportation. We use k-means clustering to organize and visualize clusters based on a calculated metric called “Availability Factor” that they have defined, and the number of transportation facilities available in each neighbourhood. Finally, they use the data at hand to create a model for multiclass classification to segregate new data into the predefined classes produced by the unsupervised learning model. The information procured in this work can be used to assess the quality of transportation available in the neighbourhoods of a location and help identify key areas for development.