Projects can be done individually or in groups. Given the variety
of interesting topics we would
like to encourage people to spread over available topics and not all
concentrate on one or two.
Data Repository
In order to explore various ways to visualize data, we are building a data
repository consisting
several kinds of data. We will have at least five kinds of data
sets in the collection:
Volumetric data
Visible woman and visible man data
Two sets of Plasma physics simulation data sets
An astrophysics simulation data set
Stock market data (currently aug, sept, and oct 1999)
NYSE Trade and Quote (TAQ) Database (user's
guide)
If you have ideas about what data to put in the repository, please let
us know.
Suggested Projects
We are actively soliciting additions to the following list. What do you
think would make a good project?
Send your ideas to li@cs.princeton.edu.
Parallel VTK for the display wall. The simplest approach would be
to run a copy of VTK on each display PC and have a master interacting with
the user and broadcast new camera information to everyone. Just like
the parallel walk-through program, it is possible to have this open-source
tool running on the display wall in days. But, after getting the
initial version running, there will be interesting performance issues such
as how to deal with huge datasets.
Similar to the parallel VTK, we can do the same with OpenDX which is also
open source. Eventually, we will figure out which one is better for
our purposes.
Visualize the stock market data. What kinds of questions would you
like to ask about the stock market? What algorithms and implementions would
you need to answer these questions and provide insight to people who like
to know about the new economy that no one seems to be able to explain?
Visualize the population data. Think of ways to visualize and analyze
the population data for various purposes. For example, if you ran
a startup company to make new-generation wireless communication gears,
which regions of would you like to offer services first?
Visualize the web server log data. Think of system design questions
you would like to ask about the web server such as access patterns and
see how to visualize the results. For example, can you determine
which web pages are important from the web data and its server log?
Visualize the genome datasets. This probably requires some understanding
of the M-Bio stuff, but it would be cool to see whether we can develop
interesting algorithms and visual metaphors to learn something non-trivial.
Visualize the a file server? How are disks allocated and being used?
Who is using the resources? What types of data are being stored,
what are the access patterns, etc? Choose an existing disk system, like
the one in the lab, and characterize it using visualization.
Collaborative, remote visualization. Some data is generated remotely.
For example, volumetric data is typically generated on a high-performance
parallel machine in a supercomputing center that doesn't do data visualization.
Some big machine probably keeps the real-time stock market data somewhere.
It may be impractical to move the data over, visualize it and make decisions.
To do data visualization remotely, what are the tradeoffs and what software
mechanisms do we need?
Participants
Stefanos Damianakis snd@cs
Adam Finkelstein af@cs
Tom Funkhouser funk@cs
Anoop Gupta
anoopg@princeton.edu
Scott Klasky
sklasky@pppl.gov
Kai Li
li@cs
Zhiyan Liu
zhiyan@cs
Robert Osada
rosada@cs
S. Jordan Parker sjparker@princeton.edu
Lena Petrovic
lenap@cs
Yilei Shao
yshao@cs
Ben Shedd
benshedd@cs
Mona Singh
mona@cs
S. Morrow Petigrew petigrew@princeton.edu
Grant Wallace
gwallace@cs
Curtis Wright
cawright@princeton.edu
