DiagramFlyer: A Search Engine for Data-Driven Diagrams
Zhe Chen, Michael Cafarella, and Eytan Adar

A large amount of data is available only through data-driven diagrams such as bar charts and scatterplots. These diagrams are stylized mixtures of graphics and text and are the result of complicated data-centric production pipelines. Unfortunately, neither text nor image search engines exploit these diagram-specific properties, making it difficult for users to nd relevant diagrams in a large corpus. In response, we propose DiagramFlyer, a search engine for finding data-driven diagrams on the web. By recovering the semantic roles of diagram components (e.g., axes, labels, etc.), we provide faceted indexing and retrieval for various statistical diagrams. A unique feature of DiagramFlyer is that it is able to "expand" queries to include not only exactly matching diagrams, but also diagrams that are likely to be related in terms of their production pipelines. We demonstrate the resulting search system by indexing over 300k images pulled from over 150k PDF documents.

Pre-print: PDF, (2.9Mb), to appear, demo track, WWW'15