An Approach to Build a Database for Crimes in India Using Twitter

Twitter is one of the most extensively used global micro-blogging site. Twitter contents have immense potential to serve the purpose of various communities. In this article, we explore the idea of using tweets regarding various crimes occurring in India to build a real-time crime database. Challenges faced throughout the work along with a naive approach on using recent natural language processing advancements, specifically in name and entity detection from texts have been highlighted. For our research work approximately 21K tweets have been collected and a prototype of database is created using our proposed method. A discussion on our findings and shortcomings in the database has been discussed and analyses on the database has been performed to arrive at certain conclusions regarding the occurrence of crimes, although such conclusions are not generalizable. To the best of our knowledge, this work is the first of its kind to aggregate tweets for building a crime database.