Authors: Weichuan Dong*, Kent State University, Xinyue Ye, Kent State University
Topics: Geographic Information Science and Systems, Quantitative Methods
Keywords: Social Media, GIS, Spatial Analysis
Session Type: Paper
Start / End Time: 1:20 PM / 3:00 PM
Room: Bayside A, Sheraton, 4th Floor
Presentation File: No File Uploaded
This study proposes two indicators describing spatial characteristics of topics in social media. Based on user relationship networks in social media, the two indicators examine local scales of topics across predefined neighborhoods such as cities and metropolitan statistical areas (MSA). The first indicator, localness of topic, describes the activeness of a topic in a neighborhood by comparing intro-neighborhood connections with inter-neighborhood connections. It calculates the density of user connections among all users in the topic within an individual neighborhood, which is an extension of clustering coefficient in social network analysis. The second indicator, spatial concentration of topic, describes the geographical influential range of a neighborhood in a topic. It calculates average distances of users from a neighborhood to their connected users both inside and outside the neighborhood. The visualization of the two indicators in geographical space reflects the activeness and influential range of the topic across neighborhoods.