Traffic Prediction in Indian Cities from Twitter Data Using Deep Learning and Word Embedding Models

Abstract

Social media platforms, such as Twitter and Facebook, have become a significant part of modern society and have provided a significant amount of text data that can be utilized to enhance text classification research using machine learning approaches. These platforms often provide more in-depth views of real-world events by correlating tweet data along with the event and its geographical location. According to research, these correlations have potential applications in a variety of real-world scenarios, including the ability to predict critical events with high precision during emergency situations. The use of real-time geolocation data helps in mapping natural and social hazards in specific areas during national calamities. In addition, the data from tweets can be used to analyse weather patterns and the sentiment or mood of people in specific localities, such as Bengaluru and Delhi. This work aims to utilize the tweets from users to report on traffic conditions in a particular area and to identify the cause of traffic congestion, which is important for understanding the severity of the problem. The goal is to modernize the traffic information system and generate crowdsourced updates for traffic, and accidents using natural language processing and machine learning approaches in Indian cities like Bengaluru, Mumbai, Delhi etc. with the involvement of Twitter users.