Date of Award
Fall 2014
Degree Type
Thesis
Degree Name
Master of Science (MS)
Department
Civil Engineering
First Advisor
Jie Shan
Committee Member 1
James S. Bethel
Committee Member 2
Hao Zhang
Committee Member 3
Ningning Kong
Abstract
With over 500 million current registered users and over 500 million tweets per day, Twitter has caught the attention of scientists in various disciplines. As Twitter allows users to send messages with location tags, a massive amount of valuable geo-social knowledge is embedded in tweets, which can provide useful implications for human geography, urban science, location-based service, targeted advertising, and social network studies. This thesis aims to determine the lifestyle patterns of college students by analyzing the spatial and temporal dynamics in their tweets. Geo-tagged tweets are collected over a period of six months for four US Midwestern college cites: 1) West Lafayette, Indiana (Purdue University); 2) Bloomington, Indiana (Indiana University); 3) Ann Arbor, Michigan (University of Michigan); 4) Columbus, Ohio (The Ohio State University). The overall distribution of the tweets was determined for each city, and the spatial patterns of representative individuals were examined as well. Grouping the tweets in time domains, the temporal patterns on an hourly, daily, and monthly basis were analyzed. Utilizing detailed land use data for each city, further insight about the thematic properties of the tweeting locations was obtained, leading to a deeper understanding about the life, mobility and flow patterns of Twitter users. Finally, space-time clusters and anomalies within tweets, which were considered events, were found with the space-time statistics. The results generally reflected everyday human activity patterns including the mobile population in each city as well as the commute behaviors of the representative users. The tweets also consistently revealed the occurrence of anomalies or events. The results of this thesis therefore confirmed the feasibility and promising future for using geo-tagged micro-blogging services such as Twitter in understanding human behavior patterns and other geo-social related studies.
Recommended Citation
Li, Yue, "Spatial And Temporal Patterns Of Geo-Tagged Tweets" (2014). Open Access Theses. 345.
https://docs.lib.purdue.edu/open_access_theses/345