Examining the potential of textual big data analytics for public policy decision-making: A case study with driverless cars in Denmark

Abstract

The simultaneous growth of textual data and the advancements within Text Analytics enables organisations to exploit this kind of unstructured data, and tap into previously hidden knowledge. However, the utilisation of this valuable resource is still insufficiently unveiled in terms of transport policy decision-making. This research aims to further examine the potential of textual big data analytics in transportation through a real-life case study. The case study, framed together with the Danish Road Directorate or Vejdirektoratet, was designed to assess public opinion towards the adoption of driverless cars in Denmark. Traditionally, the opinion of the public has often been captured by means of surveys for the problem owner. Our study provides demonstrations in which opinion towards the adoption of driverless cars is examined through the analysis of newspaper articles and tweets using topic modelling, document classification, and sentiment analysis. In this way, the research attends to the collective as well as individualised characteristics of public opinion. The analyses establish that Text Analytics may be used as a complement to surveys, in order to extract additional knowledge which may not be captured through the use of surveys. In this regard, the Danish Road Directorate could find the usefulness while understanding the barriers in the results generated from our study, for supplementing their future data collection strategies. However there are also some methodological limitations that need to be addressed before a broader adoption of textual big data analytics for transport policy decision-making may take place.