Start Date
11-8-2016
Description
Social media is a desirable Big Data source used to examine the relationship between crime and social behavior. Observation of this connection is enriched within a geographic information system (GIS) rooted in environmental criminology theory, and produces several different results to substantiate such a claim. This paper presents the construction and implementation of a GIS artifact producing visualization and statistical outcomes to develop evidence that supports predictive crime analysis. An information system research prototype guides inquiry and uses crime as the dependent variable and a social media tweet corpus, operationalized via natural language processing, as the independent variable. This inescapable realization of social media as a predictive crime variable is prudent; researchers and practitioners will better appreciate its capability. Inclusive visual and statistical results are novel, represent state-of-the-art predictive analysis, increase the baseline R2 value by 7.26%, and support future predictive crime-based research when front-run with real-time social media.
Recommended Citation
Corso, Anthony; Alsudais, Kareem; and Hilton, Brian, "Big Social Data and GIS: Visualize Predictive Crime" (2016). AMCIS 2016 Proceedings. 8.
https://aisel.aisnet.org/amcis2016/Decision/Presentations/8
Big Social Data and GIS: Visualize Predictive Crime
Social media is a desirable Big Data source used to examine the relationship between crime and social behavior. Observation of this connection is enriched within a geographic information system (GIS) rooted in environmental criminology theory, and produces several different results to substantiate such a claim. This paper presents the construction and implementation of a GIS artifact producing visualization and statistical outcomes to develop evidence that supports predictive crime analysis. An information system research prototype guides inquiry and uses crime as the dependent variable and a social media tweet corpus, operationalized via natural language processing, as the independent variable. This inescapable realization of social media as a predictive crime variable is prudent; researchers and practitioners will better appreciate its capability. Inclusive visual and statistical results are novel, represent state-of-the-art predictive analysis, increase the baseline R2 value by 7.26%, and support future predictive crime-based research when front-run with real-time social media.