Final Project
The final project for this course can consist of designing a
new visualization (that solves a problem or sheds light on a
dataset), doing usability studies to assess the efficacy of an
existing visualization, or applying known visualization techniques to
an interesting problem. Other ideas are possible, subject to
approval. Project groups can have 1-3 people.
Project Due Dates
- Oct 27: Project proposal due.
- Nov 2: (if not sooner) Receive feedback on proposals.
- Nov 14-23: Class presentations.
- Dec 12: Final writeup due. Details TBA.
The first step is to write up a proposal and have it approved by the
instructors. We may ask you to revise your proposal, most likely by
focusing the scope.
The next steps depend on what kind of project you choose to do.
If you are designing a new viz, or applying a known viz to interesting
data, then you should have an initial prototype done by Nov 23. It's
up to you how interactive your prototype is. It is best if you want
to mock it up first using paper or a very simple prototyping tool like
powerpoint or director. If you decide to implement your design,
you should first show your design to some people to get their honest
reaction before you begin coding. You do not have to code a full
working version of your design; it's up to you how interactive it is.
On Mon Oct 17, Jeff will be describing some toolkits.
The next step is to do an evaluation, with real participants, on your
design. You must have at least 3 participants and you should find
people who would be logical end-users for your design. (e.g., if you
are designing a viz for musicians, the testers should be musicians.)
See information from IS213 on
how to run informal usability studies.
If instead you are evaluating an existing visualization, you must
identify the target user population and their associated data and
tasks. You should test on least 15 participants and they should be
logical end-users for the system you are testing. You'll give a
progress report on the Nov 23rd deadline. On the final deadline
you'll report results of your study and suggestions for improvements.
Project Proposal
On Thursday Oct 27 (by 5pm), please turn in a short description of
a project proposal. Jeff and I will review it and give you
feedback. If we think it is ok, then it's done, but we may want an
iteration.
The description should include:
- Name(s) of student(s) involved
- Project goals, including what kinds of tasks the interface
containing the visualization is targeted towards.
- Discussion of related work.
- What data (if any) will be used to accomplish the goals.
- Which tools will be used to accomplish the goals (this can
change if needed).
- What steps will be required to accomplish goals.
- What kinds of results you anticipate achieving.
- What kinds of results you would like to achieve but which
you probably do not have the time or the tools for.
Project Expectations
For design projects, the design must follow good information
visualization practices, as we've discussed all semester in
class. These projects should take into account the proper use of
visualization components such as color, size, position, animation, and
so on. I will be very unhappy with improper use of visual properties.
Applications of visualization to analysis or presentation problems
should attempt to take usability issues into account, and should help
the user achieve insight on the underlying data or problem that was
not possible without the visualization. By insight I mean making the
non-visible visible, or showing trends or patterns or outliers or
missing information, or by presenting the underlying information in a
more understandable way.
If you're inventing a new kind of visualization, it might be that the
underlying results are not entirely successful. That is ok, but be
sure to follow good design principles and thoroughly discuss what did
and did not work in your design.
If you are evaluating an existing visualization, you must
identify the target user population and their associated data and
tasks. You should test on least 15 participants and they should be
logical end-users for the system you are testing.
Project Grading
- Class presentation of project results (must fit within
designated time limits) (10%)
- Quality of final writeup (40%)
- Quality of actual project (50%)