Crowdsourcing mobility insights – Reflection of attitude based segments on high resolution mobility behaviour data

• Published in: Transportation research. Part C, Emerging technologies Vol. 71; pp. 434 - 446
• Main Authors: Semanjski, Ivana, Gautama, Sidharta
• Format: Journal Article
• Language: English
• Published: Elsevier India Pvt Ltd 01.10.2016
• Subjects: Attitude based segmentation techniques; Crowdsourcing mobility behaviour; Data-driven mobility management; High resolution; Markets; Populations; Segmentation; Segments; Smart city mobility; Smartphone based mobility data collection; Support vector machines; Sustainability; Transportation; Windows (intervals); Support vector machines; Smart city mobility; Smartphone based mobility data collection; Attitude based segmentation techniques; Data-driven mobility management; Crowdsourcing mobility behaviour
• ISSN: 0968-090X; 1879-2359
• DOI: 10.1016/j.trc.2016.08.016

[Display omitted] •We explore how attitude based segments are reflected in high resolution behavior data.•We implement SVM based approach to map attitudinal segments to crowdsourced data.•The research includes 1026 users and 126,380 trips.•The success rate of the proposed approach is 98.9%. Recently, the use of market segmentation techniques to promote sustainable transport has significantly increased. Populations are segmented into meaningful groups that share similar attitudes and preferences. This segmentation provides valuable information about how policy options, such as pricing measures or advertising campaigns, should be designed and promoted in order to successfully target different user groups. In this paper, we aim to bridge between psychological, social marketing and ICT research in the field of transportation. We explore how attitude based segments are reflected in high resolution mobility behaviour data, crowdsourced via mobile phones. We use support vector machines to map eight attitudinal segments, as defined under the European project SEGMENT, to the n dimensional space defined by crowdsourced data. The success rate of the proposed approach is 98.9%. This demonstrates the applicability of the method as a way to automatically map attitudinal segments to a wider population based on observed mobility data instead of using explicit attitudinal surveys. In addition, the proposed approach can facilitate the delivery of personalised target messages to individuals (e.g. via smartphones) or at target locations where users, belonging to specific segment, are located at specific time windows since the data includes the time-space indications.