Enhancing public transportation planning through travel data analysis: a data mining application in the inference of passenger trip purpose in the Metropolitan Region of Belo Horizonte, Brazil

Resumo

Planning an efficient transport system begins with the collection of demand data. Traditionally, this data has been gathered through travel surveys conducted by interviews. These surveys involve questioning people about their trip’s starting and ending points, purpose, mode of transportation, travel duration, and other relevant details. However, this method is costly and can be somewhat inaccurate since it relies on respondents’ ability to accurately describe their journeys. With technological advancements in the transportation sector, Big Data sources have emerged as a new possibility to studying urban mobility patterns. In the public transport sector, since the 2000s the data collected by automatic fare collection systems provide a quick, accurate, and cost-effective means to estimate Origin and Destination (OD) matrices. However, a significant challenge arises when using this data source for OD matrix estimation—the lack of trip purpose information. To address this, researchers have turned to data mining techniques to inference this trip atribute. This paper contributes to the field by applying these emerging data mining approaches to infer the trip purposes of public transport passengers in the Metropolitan Region of Belo Horizonte (RMBH).