Using Twitter Data to Analyze Information Diffusion and Public Perception on Natural Hazards

Authors: Jinwen Xu*, University of Hawaii - Manoa | East-West Center, Yi Qiang, University of Hawaii - Manoa
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems, Hazards, Risks, and Disasters
Keywords: Social media, winter storm, retweet, spread, text mining, spatio-temporal
Session Type: Paper
Day: 4/3/2019
Start / End Time: 12:40 PM / 2:20 PM
Room: Marriott Ballroom Salon 1, Marriott, Lobby Level
Presentation File: No File Uploaded


Nowadays, social media are popular platforms where people communicate and share information about natural hazards and climate change. Some popular tweets can draw widespread public attention which is even beyond the significance of the actual content. As a common disaster in the U.S., Winter storms have large socio-economic impacts and were intensively discussed in social media. Analyzing the spatial dispersion of the popular tweets about the storms can help understand people’s perception and opinions towards the events and climate change. This study focuses on retweets among tweets because retweets were found to occupy a large proportion (61.4%) of total tweets. When analyzing winter-storm-related tweets, it is unavoidable to deal with retweets. Excluding retweets when computing sentiments of tweets can cause a false prediction on people's attitude. This study utilizes text mining methods to analyze the spatio-temporal variations of retweet data of Winter Storm Diego in North America in 2018. The objectives of the study include: 1) understanding the social and demographic disparities of people who respond to the hazard-related tweets. 2) analyzing the social connectivity between affected areas and non-affected areas. 3) analyzing spatial influence and spreading pattern of different types of tweets, features of largely retweeted tweets, as well as spatial clusters of users who keep concerning on winter storms. The knowledge gained from the study will give suggestions on using social media data to demonstrate spatial distribution of winter-storm-related tweets. The findings and methods of this study are applicable to other winter storms and natural disasters.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login