Using topic modelling to analyse bus route data

Loading...
Thumbnail Image

Date

Authors

Koen,H.S.
Cornelius, J.
Oosthuizen, R.

Journal Title

Journal ISSN

Volume Title

Publisher

Southern African Transport Conference

Abstract

The advent of the fourth industrial revolution and the need for connectedness have increased both data availability and quality. This data surge can also be seen in the transport and mobility industry. Anything from onboard global positioning system interfaces to vehicle trackers and wearable technology for passengers and drivers provide access to more data as an untapped source of valuable information and insights to many stakeholders. Topic modelling is traditionally used to structure and interpret text data from a large corpus of documents. In this paper, patterns in bus route data collected over several months by the onboard Global Positioning Systems (GPSs) of buses travelling in Gauteng and the Northwest province are analysed. Since topic modelling is traditionally used on text documents, the bus route coordinates had to be converted into a form readable by the algorithm. This is an ongoing project, but analyses thus far show that the most important terms per topic correspond to key nodes in city centres and points of interest where routes overlap. This information may be used in city planning to optimise the system of bus routes, terminals, and nodes. Organisations may also use this information for business development and job creation.

Description

Papers presented at the 40th International Southern African Transport Conference on 04 -08 July 2022

Keywords

d Global Positioning Systems (GPSs, Bus Route

Sustainable Development Goals

Citation