Campus Ontology-Based Correlation Detection Among Heterogeneous Data Sets

Abstract

For data-driven decision making, it is essential to build a data infrastructure that stores various data. Since various data are accumulated within organizations such as universities, companies, and local governments, integration of data in different contexts and cross-sectional analysis are issues. Knowledge graphs with a graphical structure that can flexibly change the schema are suitable for integrating heterogeneous data. In this study, we focused on a university campus as an example and proposed an ontology for various data such as lectures, buildings, purchasing, bicycle parking, and energy consumption. In particular, it has become easier to extract data across heterogeneous data collected within an organization by semantically linking dimensions with various expressions. Then, the collected unstructured data was accumulated as a knowledge graph based on the ontology, and the data infrastructure was constructed. In addition, we found several correlations through scenario-based experiment using this knowledge graph and showed the possibility that it could be applied to considering approach for university managemsent and improving the campus environment.