This document summarizes research analyzing the spatial and temporal patterns of geotagged tweets in West Lafayette, Indiana from April 11-18, 2013. The author collected over 4,000 geotagged tweets using the Twitter API and analyzed the spatial distribution of tweets on weekdays versus weekends using ArcGIS and clustering tools. Results showed different clusters of tweets by location and time of day, with more tweets originating from areas around Purdue University during weekday hours. Future work proposed analyzing tweet content and semantics to track the spread of ideas and extract broader social patterns from geotagged data.
2. 2
Introduction
• Twitter
− The most popular micro-blogging
site
− Tweets with longitude and latitude
− A gold mine for scholars in
linguistics, sociology, economics,
health, and psychology (Ghosh &
Guha, 2013)
• West Lafayette, IN
• Most densely populated city in IN
• Home of Purdue University
3. 3
Methodology
• Collect 4160 geo-tagged tweets using the Twitter
Streaming API from April 11, 2013 to April 18
• Compare the spatial distribution of geo-tagged tweets on
weekdays with those at the weekend
− Point Density tool in ArcGIS 10.1
− Clustering in Esri Maps for Excel
• Analyze the tweets on an hourly basis
11. 11
Conclusion
• Analyze the sptio-temporal pattern of geo-tagged tweets to
discover the human mobility pattern hidden behind
• Proves the feasibility of using geo-tagged tweets, in local
market research, market promotions, human mobility
analysis, and even education regulation in a “college town”
such as West Lafayette
• Future work
− Semantic analysis, topic modeling, and content analysis, aiming to
track the spread of ideas and thoughts in local area
− Framework of extracting spatio-temporal social patterns from geo-
tagged tweets in a city scale to help social researchers,
demographic surveyors, market researchers, advertisers, and
policy makers
12. 12
References
• Ghosh, D., & Guha, R. (2013). What are we ‘tweeting’about
obesity? Mapping tweets with topic modeling and
Geographic Information System. Cartography and
Geographic Information Science, 40(2), 90-102.
• Google Earth