Skip to content

Exam information

Jump to bottom
Christoph Kinkeldey edited this page Oct 26, 2020 · 39 revisions

General note

The exam consists of three parts:

  1. Interactive visualization prototype (50%)
  2. Documentation (30%)
  3. Evaluation (20%)
  • Assignments should consist of your original work. Everything else (libraries, external data, image material etc.) must be cited. Severe violation of this policy will lead to 0 points for the assignment.
  • For a successful assignment, it is more important to come up with an innovative idea, appropriate choice and preparation of data, as well as convincing visualization design rather than the perfect implementation of the visualization.

Part I: Interactive prototype (Friday, 23 Oct 2020, 1pm to 3 pm)

Design and implement an interactive visualization using the Altair library based on the COVID-19 data provided in the course as well as (optional) additional datasource(s). Deploy the visualization so that it is publicly available on the web (e.g. use Github Pages).

Deliverables

  1. Link to working prototype on the web
  2. Presentation slides

Project presentation

Each group will give a presentation of about 10 minutes.

  • We will have live project presentations with slides in English.
  • This includes the demonstration of the prototype and the design decisions.
  • Please divide up the presentation to make sure that all group members actively take part.
  • After the presentations we will have 5 minutes for questions.

Grading

During the presentation we will evaluate the following aspects:

  1. Question(s) addressed (clarity and quality)
  2. Data preparation (choice of external data source, data wrangling, handling of missing data)
  3. Design (visual encoding / interaction)
  4. Presentation (structure, clarity, timing)

After the presentation we will have a closer look at:

  1. Novelty (does a very similar solution exist?)
  2. Choice and preparation of data sources
  3. Visualization and interaction design
  4. Interface layout and linking of views

(+) Additional points can be earned through:

  • Creative use of novel visualizations
  • Advanced interaction techniques
  • Effective multi-view coordination
  • Thoughtful and elegant graphic design
  • Engaging exploration experience

(-) Point reduction can result from:

  • Solutions very similar to existing ones
  • Ineffective visual encodings
  • Lack of meaningful interaction techniques
  • Confusing interface design or unhelpful interactions

Part II: Documentation (6 Nov 2020, until 11:59 pm)

Deliverables

  • Project documentation as README.md in your GitHub repository master branch (~1500 words)
  • Screencast embedded in the documentation

Project documentation

Describe the concept of your interactive visualization in your own words. Do not forget to describe your data sources incl. their license information. Mention possible improvements for your project (if you had more time). If you used other libraries or frameworks other than Altair please explain why briefly.

In addition, for a formal description of the design decisions, use the “Four Nested Levels of Visualization Design” by Munzner:

a. Domain problem characterization: Describe the domain situation, i.e., what is the group of target users, their domain of interest, their questions, and their data.

b. Data / task abstraction: Formulate the tasks in a domain-independent vocabulary. How did you prepare (aggregate, filter...) the data to support the tasks?

c. Visual encoding / interaction design: Describe the visual encoding and why you decided for it. What interaction types did you use and why? (be precise and try to cover all details in this part)

d. Algorithm design: How did you make sure that the computational complexity of your solution is appropriate? What is the bottleneck with respect to performance?

Screencast

The screencast should, different than the project presentation, give a short introduction in the visualization and serve as a teaser as well as a user guide. It should be around 3 minutes in length.

Part III: Evaluation (6 Nov 2020, until 11:59 pm)

Evaluate the design of the "next" group:

  • Group 1 evaluates group 2
  • Group 2 evaluates group 4
  • Group 4 evaluates group 5
  • Group 5 evaluates group 6
  • Group 6 evaluates group 1

Deliverables

  • Documentation as EVALUATION.md in your GitHub repository master branch (~500 to 1000 words)

Try out the online prototype. From the presentation video, get more information on the target group etc.

  1. Use the Cognitive Walkthrough method to walk yourself through the visualization from the perspective of a person from the target group using this visualization. Write a short story about how a person from the target group uses the visualization and what question he/she can answer with it. Does the person run into problems (e.g., reading the visualizations, using the interface etc.)?

  2. As formal validation, address the threats mentioned in the Munzner nested model for the first three categories (domain, data/task, and visual encoding / interaction).

  3. What possible improvements do you suggest as a result of the validation?

Clone this wiki locally