Database Reference
In-Depth Information
some terms augmented by selected papers, and other terms I have used only
research papers. I have settled on having students read one or possibly two
research papers for each class. Typically, the paper is an important one for that
topic or it is a good overview of the issues involved. When I have assigned more
papers than this, I find that the students often do not adequately prepare and
read all the papers. To cover more recent research, I typically select two or
three recent articles on the topic of the day and I assign two or three students
who must recap and describe their particular paper's key ideas to the class in
less than five minutes. I believe that experience giving presentations like this is
important and valuable to the students. All of my lecture slides can be found at
the course website and in 2007 I created in-studio video versions of each lecture.
These videos can be found at the website http://vadl.cc.gatech.edu .
I use a number of relatively small homework assignments in the course,
along with one larger homework and a group project. I will frequently employ
a midterm or final exam as well. The small homeworks often involve a visual-
ization design exercise (on paper) given a data set. Of course, such assignments
do not engage the interactive component of information visualization that is so
important, so they are fundamentally limited.
The larger homework assignment is a commercial tools critique. Students
are given five example datasets and asked to choose the two that they find
most interesting. Before using any systems, the students examine the datasets
and generate questions about them. Next, the students use a few information
visualization systems to explore the data and try to answer those questions.
I also alert the students to note any serendipitous findings that occur during
exploration. Finally, the students must write a report in which they critique the
different systems used, the visualization techniques each employs, and whether
the systems led to insights and discoveries. I have used systems such as Spotfire,
SeeIt, Advizor, Eureka (Table Lens), InfoZoom, InfoScope, and Grokker over
the years. I find this assignment to be extremely valuable to the students as it
allows them to gain hands-on experience with sophisticated systems and shows
them how visualizations can (or cannot) be helpful in analysis and exploration.
This particular assignment even led to an interesting research contribution
by my group. We studied the analytic queries generated by students over many
years of the course and clustered these inquiries into different low-level analytic
tasks that visualizations may assist. Our taxonomy of these tasks was presented
at the 2005 Symposium on InfoVis [1].
I also employ a group project in the course in which students design and build
a visualization system for a particular problem and data set. Teams of three or
four students work together for most of the term and find a client with a data
analysis problem or they simply choose a data set and envision the kinds of ana-
lytic queries that one would expect on it. The students explore different visualiza-
tion designs, then they choose one to implement. In the past, student teams have
often chosen to work on the contest datasets from the IEEE InfoVis or VAST
Conferences. In fact, student teams from my course have won these contests on
multiple occasions or have had competitive entries in the contests [20,54]. Group
projects have even led to full papers at the InfoVis Symposium as well [12].
Search WWH ::




Custom Search