Date of Award
Master of Science
Budhendra L. Bhaduri
Bruce A. Ralston, Hyun Kim
With society’s increasing participation in social media, scientists now have access to new sources of data that reflect our daily activities in space and in time. Such data are plentiful and, more notably, at an unprecedented granular level. The ability for users to capture and express their geolocation through their phones’ global positioning system (GPS) or through a particular location’s hashtag or Facebook Page provides a great opportunity for modeling spatiotemporal population dynamics. High resolution population models and databases for episodic special events can be extremely useful for enhancing emergency management and response. This research assesses the feasibility of improving a special event population distribution and dynamics model, namely Oak Ridge National Laboratory’s LandScan USA, using data from social media. Specifically, analysis is across a 24 hour period for a number of football game days associated with the University of Tennessee, Knoxville during the 2013-2014 season. Data from two popular social media platforms, namely Twitter and Facebook, were used to analyze possible patterns of population distributions around the university’s football stadium. Spatial autocorrelation was measured and calculated using Global Moran’s I and the Local Indicator of Spatial Association (LISA) test to support and build confidence of the tweet and check-in data. Overall, data from social media were found to be beneficial for improving high-resolution population distribution datasets, such as LandScan USA. However, long term collection and analysis of social media data are necessary for ensuring sustainability and predictive capacity of such data in modeling near real-time population dynamics for special events.
Sims, Kelly Michelle, "Integrating Social Media in the Development of a Special Event Population Dynamics Model. " Master's Thesis, University of Tennessee, 2014.