This is a project website for semi-automated analysis of scientific article references conducted in spring 2019
The project allows the user of the system (working as admin, not as any user) to add publication information to the system by means of PDF-files. The system then uses automated logic to find the reference section in the publication, then parses these references into structured information, extracting authors, publication year and title.
Since PDF is a layout format and does not have a rigid structure, all layout cannot be matched by the parser. However, with scientific articles being quite "un-creative" in terms of layout, the results are pretty good.
This website currently has 16 publications added to it. The system has automatically parsed the references of those publications and is able to compute key information based on this data (more than 700 references and about 1200 authors). Below, one can click a link to see figures based on this group of primary reference.