A Methodology for Aligning Categories from Open Government Data Portals to a Comprehensive Set of Categories

Nowadays, we can find Open Government Data Portals (OGDPs) from different governmental organizations. The increasing number of OGDPs brings opportunities for integrating these data. The vast majority of OGDPs distribute the data into different categories, and each portal uses its own set of categories. Therefore, due to the lack of homogeneous guidelines in data management, interconnecting data is flawed. Data integration is crucial in improving the functioning and planning of open government data portals. One possibility for integrating data from different portals is using the categories associated with datasets. Putting similar datasets in the same category can turn data integration easier. We propose a methodology for constructing a Comprehensive Set of Categories (CSC) extracted from different OGDPs and aligning different categories to this minimal and comprehensive set based on semantic similarity. We carried out an exploratory analysis on 100 portals of densely populated American cities using the categories collected in these portals in 2017. Our approach allowed us to align more than 80% of the collected portal categories to the minimal set according to more than 3 out of 6 similarity measures, which is promising for open data integration.