Visualizing Paper Genealogy
Justin Matejka is a Research Assistant at Autodesk Research, a firm dedicated to innovation in digital prototyping, visualization, and simulation tools and techniques toward advancing design. As part of the Visualization and Learning projects, Matejka created Citeology, a complex interactive visualization of the relationship between research papers through citations. It is accompanied by a paper as well, which was recently submitted to alt.chi for the CHI 2012 conference on human computer interaction research and design.
If one considers a research article in a genealogical sense, the papers which an article referenced could be considered the article’s “ancestors” or “parents” and the papers which referenced the target article could be considered “descendants” or “children”… Such information can be useful when tracing the history of a piece of work or trying to find related articles. It is however generally presented textually and there is no way to look at multiple generations of ancestors or descendants or get a feel for the overall network of the corpus.
Thus, he created an interactive graphic that allows a viewer to get this feel for the overall network. Using all 3,502 papers published at the CHI and UIST Human Computer Interaction conferences between 1982 and 2010, Matejka represented all 11,699 citations made from one article to another on a series of curved lines. Past articles referenced in a given article are depicted in blue, and future articles which reference it are in red.
Clicking on a single paper allows you to see its specific genealogy. When one paper is selected, hovering over other papers allows you to see the shortest connected between the selected paper and the hovered paper. Papers can also be looked up by title or author through a search bar.
The project is especially interesting because of its potential to visualize the specific internal connections between papers and also the broader corpus of a work, as Matejka says. The applet could be a useful tool for educational research and is also an ideal dataviz project to demonstrate the use of data art in education.