In this study, we demonstrate how the information provided by individuals on social media can represent some aspects of their behavior to predict criminal activities and intentions in societies. The problem discussed in this study is finding a model that can analyze social media posts to infer the intentions and feelings of the publisher behind those posts to predict the probability of committing a crime. This is a preventive technique that can be used to monitor individuals or organizations who have a behavioral pattern that can be inferred as criminal intent. To help detect and predict criminal activities, we observe the use of data mining followed by sentiment analysis on Twitter. This well-known online social network enables users to post small texts, aka "tweets," that are up to 280 characters in length each. In the U.S. was the main That of the study and data collection. First, the targeted tweets were collected according to geographical and keyword-based filters. Then a sentiment analysis was applied to analyze the crime intensity in specific locations. A correlation was found between the collected tweets related to criminal activities and the crime rates in the corresponding cities. Furthermore, another analysis study that we based in the United Arab Emirates found out that the quality of tweets is lower than the tweets in the United States due to the number of spam tweets. The lack of coloration between tweets and crimes committed on the ground due to laws prohibiting sharing information of crimes from police departments.
Professional Studies (MS)
Department, Program, or Center
Graduate Programs & Research (Dubai)
AlSubousi, Rashed Khalifa, "Feasibility of Twitter sentiment analysis in predicting crime in the UAE" (2021). Thesis. Rochester Institute of Technology. Accessed from