max planck institut
informatik
informatik
Mike Sips (VIS) Arturas Mazeika (Data Mining)
|
Graphical inventions and displays are an important external aid that help the human analyst to reveal hidden patterns in large and complex data
spaces. In this class, students will get acquainted with latest research in exploratory data analysis and interactive visualization. The students
will implement an own visualization technique. The goal of this class is to develop a clever and easy to use visualization that makes a difficult
problem easy to solve; rather implementing a complex visualization tool.
The students will have to team-up in groups (up to three students) to implement a visualization tool. The final grade of the course will be
determined based on (a) the groups ability to demonstrate the merits of their implemented visualization technique to solve the problem at hand (80
%), and (b) the knowledge of each group member of basic concepts in exploratory data analysis and interactive visualization (20%).
Groups are highly encouraged to submit their own project proposal, but each group can choose one of the project proposals listed below. In each group
only students that are interested in addressing a particular problem should team-up.
Our project
proposals describe research problems considered currently "hot" in the visualization and data mining communities. Each project proposal has a
short problem statement and a suggested plan of work. The students will need to read and digest the given papers to get experienced in the
state-of-the-art. Each group will need to propose an idea of a potential visualization technique that is either (a) an extension of the
state-of-the-art, or (b) might be a new and clever idea.
The class will meet twice a week for a lecture (2h), and laboratory work (4h). The lecture will discuss basic concepts of visualization, exploratory
data analysis, and visual analytics, and latest research topics as well. The laboratory work includes discussion of project milestones, design
decisions, and feedback. The laboratory work is meant to keep the groups on track to implement a clever solution.
|
There are no prerequisites for the class, and the class is open to students at all levels. bachelor. However, a basic working knowledge of a graphics API (e.g. OpenGL, Java2D/3D) will be useful. The programming project can be developed using any suitable language. While these APIs, applications and languages will not be taught in the class, many introductory tutorials at the level required for the class are available on the web. The class language is English. The class has 9 credit points .
|
Groups are highly encouraged to submit their own project proposal. It is also possible to chose a project proposal listed below. All groups will be
required to present their proposal for the programming project in the first laboratory assignment. There will be room for discussions and feedback.
Note all groups (own + selected) are required to submit a research proposal not later than the announced deadline. We will send a notification of
acceptance for both selected and own proposals.
Each proposal should include:
|
Deadline: postponed to October 31 2009
(a) Information Visualization
|
(b) Interaction Techniques
|
(c) Exploratory Data Analysis
|
In addition to the final exam, students may optionally participate in a visualization competition. While grades for the projects are based mainly on "usability merit", entries in the competition will be judged on technical merit, compelling visual encoding and originality. The winner will be selected by a panel of visualization experts from industry and academia. The competition will take place at the end of the semester.
|
Data used in the clases: data.tgz (76MB), data generation code (2.2MB)