GeoFlux: Hands-Off Data Integration Leveraging Join Key Knowledge

Data integration is frequently required to obtain the full value of data from multiple sources. In spite of extensive research on tools to assist users, data integration remains hard, particularly for users with limited technical proficiency. To address this barrier, we study how much we can do with no user guidance. Our vision is that the user should merely specify two input datasets to be joined and get a meaningful integrated result. It turns out that our vision can be realized if the system can correctly determine the join key, for example based on domain knowledge. We demonstrate this notion by considering a broad domain: socioeconomic data aggregated by geography, a widespread category that accounts for 80% of the data published by government agencies. Intuitively two such datasets can be integrated by joining on the geographic unit column. Although it sounds easy, this task has many challenges: How can we automatically identify columns corresponding to geographic units, other dimension variables and measure variables, respectively? If multiple geographic types exist, which one should be chosen for the join? How to join tables with idiosyncratic schema, different geographic units of aggregation or no aggregation at all? We have developed GeoFlux, a data integration system that handles all these challenges and joins tabular data by automatically aggregating geographic information with a new, advanced crosswalk algorithm. In this demo paper, we overview the architecture of the system and its user-friendly interfaces, and then demonstrate via a real-world example that it is general, fully automatic and easy-to-use. In the demonstration, we invite users to interact with GeoFlux to integrate more sample socioeconomic data from data.ny.gov.