Analyzing Tweets to Understand Factors Affecting Opinion on Climate Change | Shiv Nadar University
Enquire Now
Apply Now
Undergraduate Admissions – Information Session ScheduleInternational AdmissionsM.Tech / PG ProgramsPhD Admission Spring 2021MBA (Executive) AdmissionsData Sciences and Analytics for Business (DSAB)

Analyzing Tweets to Understand Factors Affecting Opinion on Climate Change

Research
24 Jul 2021

Citation
Mohith, S., Jackson I. Jose, Sonia Khetarpaul, and Dolly Sharma. "Analyzing Tweets to Understand Factors Affecting Opinion on Climate Change." In Databases Theory and Applications: 32nd Australasian Database Conference, ADC 2021, Dunedin, New Zealand, January 29–February 5, 2021, Proceedings 32, pp. 99-110. Springer International Publishing, 2021.

Abstract
Climate change is a topic that is frequently debated on social media. A vast majority in the debate cite scientific evidence to recognize the existence of a man-made climate change and its impacts on environment as well as society. The opinion of the masses is critical to dealing with various issues arising due to climate change, such as global warming. In this work, we study people’s opinion on climate change and analyze the data to identify the common topics which garner discussion. Our dataset includes crawled tweets containing the words ‘‘climate change’’, ‘‘global warming’’ and other related keywords. Our aim is to analyze the dataset, explore the popular belief of a region and then derive the possible explanation in terms of different factors. This analysis could help us in determining the extent to which different factors affect people’s opinion. Multiple models for sentiment analysis were tested, and the classification for the same is done as a two-step process. First – (i) classify news and opinion-based tweets, and then (ii) classify the opinion-based tweets as being negative, positive towards the subject. Topic modelling is performed to find the topics which gather attentions of people. After analyzing about 350k tweets, various sentiment-based and topic wise distributions are visualized and studied.

Click here for details

Directions