01657489

Component

Social media have been recently employed as a source of information for event detection, including traffic congestion and accidents. Twitter, one of the social media platforms, is the most popularly used one for such applications due to the availability of an application program interface (APIs). Twitter allows users to share information in 140 characters and gives the location of the incident and context of the incident. This study explores the possibility of using Twitter to obtain traffic information under Indian conditions, taking Chennai and Bangalore as case study. The system fetches tweets using Twitter Streaming API as well as Search API, pre-processes tweets using pre-filtration and text-mining techniques and finally performs binary classification of tweets using lexicon based approach and supervised machine learning (ML) based approach. Supervised ML-based approach was used also to classify tweets into finer classes to know the reason behind the congestion.

This paper was sponsored by TRB committee ADA60 Standing Committee on Public Involvement in Transportation.

18-02275

English

7p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; Maps; References; Tables

Advanced traffic management systems; Automatic data collection systems; Data mining; Incident detection; Incident management; Machine learning; Social media; Traffic data; Traffic incidents

Twitter

Text mining

Bangalore (India); Chennai (India)

Data and Information Technology; Highways; Operations and Traffic Management

Transportation Research Board Annual Meeting 2018 Paper #18-02275

TRIS, TRB, ATRI

Jan 8 2018 10:33AM