A query log analysis of dataset search

Abstract

Data is one of the most important digital assets in the world and its availability on the web is increasing. To use it effectively, we need tools that can retrieve the most relevant datasets to match our information needs. Web search engines are not well suited for this task, as they are designed primarily for documents, not data. In this paper, we present the first query log analysis for dataset search, based on logs of four national open data portals. Our aim is to gain a better understanding of the typical users of these portals and the types of queries they issue, and frame the findings in the broader context of dataset search. The logs suggest that queries issued on data portals differ from those issued to web search engines in their length and structure. From the analysis we could also infer that the portals are used exploratively, rather than to answer focused questions. These insights can inform the design of more effective dataset retrieval technology, and improve the user experience of data portals.