Understanding Travellers’ Satisfaction with and Attitudes Towards TfL Bus Service by Twitter
Weijia Chen, Newcastle University, Amy Weihong Guo, Newcastle University, Phil Blythe, Newcastle University
This paper reports an innovation method of integrating an online survey data and data from Twitter to understand the utilisation of social media in changing travellers' travel behaviour and their attitudes improvements towards TfL bus service.
Social media sites provide many-to-many interaction opportunities for transport operators to communicate with large numbers of travellers simultaneously in real-time. An increasing number of transport agencies throughout the world have adopted social media for practice. At present, most of the studies have investigated the utilisation of social media in general practice, but the research of social media in changing travellers’ travel behaviour to become more sustainable has not been fully investigated.
This paper reports a recent study that utilised an innovation method by integrating online survey and data from Twitter to understand the utilisation of social media in changing travellers’ travel behaviour and their attitudes improvements Transport for London (TfL) bus services. Online survey data was collected from 25th November 2014 to 25th January 2015. 512 data were collected in total, in which 481 were analysed as valid data. Data from Twitter were collected from 21st July 2014 to 25th January 2015.
After analysing 481 online survey data, it was found that 47% of surveyed bus passengers stated that TfL bus Twitter service improved their attitudes towards TfL, and 13% of all respondents increased their bus trips based on the two-way communication with TfL by Twitter. In addition, 76% of all respondents claimed that it is easier to connect with TfL by Twitter than other media channels, such as telephone and email, in which 75% of those passengers connect more frequently with TfL than before. Furthermore, 55% of all respondents reported that the replied tweets from TfL bus Twitter account solved their enquires and issues, and faster than traditional medial channels as well.
The findings indicate that 66% surveyed bus passengers changed their behaviour, such as changed departure time or journey rout, based on the communication with TfL by Twitter. Furthermore, research findings provide insights into gender differences in attitudes and behaviour changes based on TfL bus Twitter service. It is suggested that it is easier for females than males to change their travel behaviour and improve their attitudes towards TfL through the Twitter services that are provided by TfL.
After analysing the online survey data, data from Twitter was combined with online survey data to investigate whether the sentiment analysis of tweets as a technique can be applied in transport to understand travellers’ attitudes towards TfL bus service. It is very common that more than one attitudes can be included within one passengers’ tweet message. For example, “Thanks @TfLBusAlerts - no joy unfortunately :(” This tweet includes both negative and positive attitudes. Therefore, in order to understand the complex attitudes from travellers, SentiStrength 2.0 was used to analyse the tweets data.
SentiStrength 2.0 is a lexicon-based classifier that integrates linguistic rules in calculating sentiment strength in short information English text. For each text, the SentiStrength 2.0 output is two integers: 1 to 5 for positive sentiment strength and a separate score of -1 to -5 for negative sentiment strength. 1 signifies no sentiment and 5 signifies strong sentiment of each type.
After calibration the SentiStrength 2.0, both positive and negative scores of 20,000 tweets were quantified to identify travellers’ attitudes towards TfL bus service. In addition, the results of sentiment analysis score were compared to the results of survey data that was designed to ask travellers’ satisfaction towards TfL bus service based on five-point likert scale questions from very dissatisfaction to very satisfaction. The findings provided insight into the sentiment analysis, and it can be seen as a supplementary tool to the traditional customers’ satisfaction survey.
This paper provides the evidence and confidents for transport stakeholders in utilising social media platforms to changing travellers’ travel behaviour and improving their attitudes toward transport operators’ services. The full results will be elaborated in the final paper and presented at the conference if accepted.
Association for European Transport