Visualization of figure placement in (visualization) papers



Figures are a crutial element in scientific papers, needed for proper explanation of the matter. It is important to use figures (images and tables) effectively to convey the message, therefore most of the emphasis is (rightly) on their content. In this project you will however explore the role that position and size of a figure play in their usage. You will look at papers from the field of visualization itself.

Your task will be to create a visualization system that fetches papers published in the IEEE Vis conference and creates an aggregated visualization that reveals patterns and trends in how authors use figures.

Using such visualization, these questions could be answered:

  • Do the patterns of figure placement differ across SciVis, InfoVis, and VAST tracks?
  • Is there an "unwritten" best practice when it comes to placing figures in the paper?

This topic is more suitable as a student project (because of it's emphasis on implementation) but can be used as a topic for bachelor's thesis as well.


  • loading of a large numer of papers (in the PDF form), ideally in a semi-automatic manner
  • parsing the PDF for figure size and placement
  • visualization where the placements from each individual publication are aggregated (imagine a heatmap or similar)
  • evaluation, discussion of the findings


  • Knowledge of English language is required
  • Knowledge of C++ and Qt is advantageous, but not necessary (technology choice is on the student)


The project should be implemented as a standalone application, desktop or web-based (to be discussed).


For more information please contact David Kouřil (