Application or Semester Project Ideas

From Cmsc734_08
Jump to: navigation, search

Application or Semester Project Ideas

  1. Prof. Susan Moeller- Journalism, will have some projects
  2. Darya Filippova (in our class) – Extend Basketlens for text analysis
  3. Catherine Plaisant will have a project for National Cancer Institute – patterns of cooperation among 650 organizations.
  4. CS Grad student is ready to work with you: Galileo Namata [namatag@cs.umd.edu]- Improve last year’s class project on DualNet.
  5. Beth Noveck (New York Law School): Apply visualization to patent review process:
  6. Catherine Plaisant and grad student David Wang:
  7. Scott Gilkeson, State of the USA – public info on key indicators
    1. Comparing indicators - simple, easily understandable but powerful ways to compare two or more indicators to make a point that would not be obvious otherwise (to include methods for selecting indicators to compare and tools for visualizing the multiple indicators)
    2. Policy scenarios - choose a policy issue and present both sides of it backed up by data supporting each viewpoint
    3. Social networking - evaluate social networking tools as applied to policy discussions by posing an issue and working through it using data (probably on a site like ManyEyes, Swivel, or possibly Data360 which already have some social networking capability).
  8. Prof. Vibha Sazawal projects:
    1. Info Vis for debugging. I am particularly interested in applying results from medical informatics regarding medical diagnosis to bug diagnosis.
    2. Learning a new programming language given knowledge of another programming language. I see this as an ontology mapping problem.
  9. Catherine Plaisant: Improve the Timeline for HCIL projects:
  10. CLIS Prof. Jimmy Lin: I used Social Network Analysis (SNA) to visualize related document networks, i.e., documents connected by content similarity links. We considered graphs on the order of hundreds of nodes, but what about: 17 million nodes! all of MEDLINE... As part of the cloud computing course, a team will be working on these large networks, doing things like computing PageRank, finding communities of related papers, etc. See:
  11. Neutrino experiment (Troy Straszheim)
    • Icecube - International physics collaboration
    • Detector at magnetic south pole instrumenting one cubic kilometer of Antarctic ice that looks for Cherenkov radiation from neutrinos that have come from outside our galaxy and already passed through the earth. There is little agreement on the representation of the data but are generally shown as time series. Current implementations include a Qt-based visualization tool.
  12. Searching and sorting versions
    • The Versioning Machine (http://www.v-machine.org/) was built to provide an application that displays images and encoded transcriptions of different versions of manuscript texts all at once for comparison and analysis. The application, however, does not currently support any mechanism for searching or sorting these displays or for facilitating comparison across or between document sets. In particular, this project will be focusing on an iteration of the Versioning Machine that is being used within the Baroness Elsa von Freytag-Loringhoven Digital Library, which Tanya Clement is augmenting with new poems one of which is primarily image-based. Some questions might included : can changes in lines that are out of parallel be visualized? What about changes in prose versus poetry? What about similarities or differences across document sets? Line or word collocation? Is there a visualization that would allow a user to know what is in the library in general and how these objects all relate?
    • With grad student Tanya Clement
  13. Named entity extraction manager
    • The goal of this project would be to address a common problem in information management. Given that named entity extraction is so poor and that such processes require so much “clean-up,” is there a way to create an environment in which a user can facilitate the process of extraction while providing a model for machine learning? In other words, instead of creating a tool for cleaning named entities, create a tool for creating those entities and managing how they are related by visualizing relationships the user determines are links between and across sets of names. There are three possible data sets that may be available for consideration in this project: a 900-page novel and a much shorter, short-story collection (The Making of Americans and Three Lives by Gertrude Stein), a literary biography titled Baroness Elsa: Gender, Dada, and Everyday Modernity--A Cultural Biography by Irene Gammel, or an archive of literary correspondence with Thomas MacGreevy (for examples see http://www.macgreevy.org/).
    • With grad student Tanya Clement

Remember: project ideas should have a real user and one where the results of your work will live on beyond the semester.