Ontology-based Instance Matching for Geospatial Urban Data Integration

Abstract

To run a smart city, data is collected from disparate sources such as IoT devices, social media, private and public organizations, and government agencies. In the US, the City of Chicago has been a pioneer in the collection of data and in the development of a framework, called OpenGrid, to curate and analyze the collected data. OpenGrid is a geospatial situational awareness platform that allows policy makers, service providers, and the general public to explore city data and to perform advanced data analytics to enable planning of services, prediction of events and patterns, and identification of incidents across the city. This paper presents the instance matching module of GIVA, a Geospatial data Integration, Visualization, and Analytics platform, as applied to the integration of information related to businesses, which is spread across several datasets. In particular, we describe the integration of two datasets, Business Licenses and Food Inspections, so as to enable predictive analytics to determine which food establishments the city should inspect first. The paper describes semantic web-based instance matching mechanisms to compare the Business Names and Address fields.