Description
Whereas ad hoc single domain Big Data inquiry is successful, observation of a multi-domain GIS artifact needs consideration. A GIS solution for multi-domain data analysis must provide visualization and overt statistical analysis tools, e.g., regression capabilities of constituent data streams, in order to enable large-scale dataset processing and evaluation. Such guidelines direct inquiry and creation of a robust GIS artifact considering a social media tweet corpus and a domain specific crime dataset. The tweet corpus is operationalized via natural language processing treatments and used in GIS artifact construction and evaluation. Although results are not statistically significant and visualizing crime data is not novel, learning how to combine the two in predictive ways via GIS is. As such, extensions and possible future work support social media natural language processing techniques and Big Data processing for predictive crime-based incident interactions as front-run by real-time social media analysis.
Recommended Citation
Alsudais, Kareem and Corso, Anthony, "GIS, Big Data, and a Tweet Corpus Operationalized via Natural Language Processing" (2015). AMCIS 2015 Proceedings. 34.
https://aisel.aisnet.org/amcis2015/BizAnalytics/GeneralPresentations/34
GIS, Big Data, and a Tweet Corpus Operationalized via Natural Language Processing
Whereas ad hoc single domain Big Data inquiry is successful, observation of a multi-domain GIS artifact needs consideration. A GIS solution for multi-domain data analysis must provide visualization and overt statistical analysis tools, e.g., regression capabilities of constituent data streams, in order to enable large-scale dataset processing and evaluation. Such guidelines direct inquiry and creation of a robust GIS artifact considering a social media tweet corpus and a domain specific crime dataset. The tweet corpus is operationalized via natural language processing treatments and used in GIS artifact construction and evaluation. Although results are not statistically significant and visualizing crime data is not novel, learning how to combine the two in predictive ways via GIS is. As such, extensions and possible future work support social media natural language processing techniques and Big Data processing for predictive crime-based incident interactions as front-run by real-time social media analysis.