Intelligent Graph Reader

Client: Epistora

Main client: Nilu Satharasinghe <nilu@epistora.com>

Additional contact: Vivian Chan <vivianchan2@gmail.com>

Google Scholar is a good way to find scientific results, but it only reads the words - not actual data. The goal of this project is to create document analysis algorithms that can automatically identify and describe graphs of scientific data - something like "find a graph where blood pressure decreases with age". Users should be able to feed in complete HTML or PDF documents, which are automatically parsed for graph content, and stored in an archive that supports sophisticated data queries and comparisons between graphs in different documents.