AAAI Publications, Workshops at the Twenty-Eighth AAAI Conference on Artificial Intelligence

Font Size: 
Dexter: Plugging-n-Playing with Data Sources in your Browser
Abhijeet Mohapatra, Sudhir Agarwal, Michael Genesereth

Last modified: 2014-06-18


We present Dexter, a browser-based, general purpose data exploration system for end-users. Dexter enables end-users to easily query across multiple Web-accessible heterogeneous (semi-) structured data sources with higher expressivity than that is usually directly supported by the sources. A novelty of our approach lies in the client-sided evaluation of end-user queries. Our query evaluation technique exploits the querying capabilities of the sources and communicates directly with the sources whenever possible. Dexter-Server, the server-sided component of Dexter, merely acts as a proxy for accessing sources that are not directly accessible from a Dexter-Client. Dexter also supports organizational internal and personal data sources while respecting end-users' security and privacy. We present the results of our evaluation of Dexter or scenarios that involve querying across data about the U.S. Congerss, the U.S. Code, and feeds from popular social networks. We discuss the applicability of Dexter for data about cities.


Small-Data; Web APIs; Querying; Datalog; Privacy; Browser-based; Data-Exploration

Full Text: PDF