EXPLORING AND COMPARING THE QUALITY OF PUBLIC TRANSPORTATION SERVICES FROM SOCIAL MEDIA: CASE STUDY OF GREATER JAKARTA

• Main Author: Marantino Napitupulu, Hendrik
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/51743
• Institution: Institut Teknologi Bandung

Personal vehicles were allegedly the main cause of air pollution and traffic jams in Jakarta. Private vehicle users in Jakarta are very high at 77% compared to public transport users. Switching to public transportation is one way to tackle the two main problems. To suggest people to switch from their private vehicles to public transportation is to provide availability of transportation network and also convenient. The purpose of this study is to see how was the quality of service by the two largest public transportation modes in the Jakarta which are Bus Rapid Transit (BRT) and Commuterline (KCI) with approach of society status on their social media. The method used in this research are sentiment analysis which will be complemented by topic modeling. Data collection was carried out for 6 months. From the results of the model, it was found that in the sentiment analysis the F1- score was 74% for both modes, while the coherent scores in topic modeling was 0.5002917973934803 for BRT and 0.4760793308842305 for KCI. From the above reliability of the model that we have made, most of the users from the two bigest modes of transportation in Jakarta still perceive their services poorly. The biggest problem with the BRT mode is the case of vehicle entry its lane, whereas in KCI mode is the frequency of trains waiting for their turn to enter the station, which is interpreted from keywords that come out of the results of the analysis where there are many other important words. The results of this study are very useful for the operators of transportation service providers in understanding the needs of passengers, as well as for local governments to find out the community's response to the policies they make. For further research, can be applied with other modes of transportation with a longer time and connect the sentiment analysis to topic modelling, therefore from each topic we can see the public sentiment and make rating by topics.