A4: Low-Fidelity Design

Due date: May 22 (23:55)

Experts from various professions are looking for skilled visualization experts who can evaluate a broad set of climate change data and build an informative tool with which to perform various tasks. Imagine that you were hired to build this visualization tool.

Data

The climate change data is composed of multiple datasets. The complete collection of the datasets can be found here. The folder contains a summary file, “0_Indicators_Summary.csv,” that describes all datasets. The main focus should be on the climate datasets (#13 – 15 and #21 – #26). However, you are allowed, and we encourage you, to use the other datasets (#01 – #14, #16 – #20) as well. Pick at least 3 different datasets from the pool, where at least one of them has to come from datasets #13 - #15 or #21 – #26.

Study the datasets and their characteristics before proceeding with the tasks below.

Users

We provide you with four expert user groups for the tool. Please pick one user group of the first two and one of the second two, and the corresponding task from each of the selected user groups. We expect you to come up with at least two more tasks for each user group on your own so that you end up with two user groups and at least three tasks per user group.

Pick one of the following:

• Service providers for greenhouse emission compensation
  • Task: Investigate possibilities to mitigate climate change effects which can be offered to frequent flyers who want to decrease their carbon footprint, e.g., planting trees, direct air capture, disaster relief.
• Activists from WWF
  • Task: Investigate the relative impact of climate change on natural habitats in different parts of the world regarding, e.g., natural disasters, sea level, or forest cover

Pick another one of the following:

• UN commission on climate change
  • Task : Demonstrate which countries are affected most by climate change to debate strategies
• Data journalists (e.g., Standard or Presse)
  • Task: Demonstrate how climate change indicators in Austria compare to other countries

Directions

We expect you to present two different dashboards, one for each selected user group. Each dashboard must contain at least 4 different views and 3 different visualization types (scatterplot, histogram, map, etc.) for different facets of the datasets. For each design, please report on how the visualization or interaction contributes to a task outlined earlier. Refer each plot explicitly to one (or more) of the tasks you identified before.

We recommend using paper prototypes for this phase. Paper prototypes have the advantage that you are only limited by your imagination as to what is possible. If you have a novel visual encoding then feel free to include it!

Submit a PDF report containing the following (max. 8 pages):

• Describe your two types of users in more detail
• Characterize the dataset in detail (interesting features, missing values)
• Perform a detailed task analysis: Describe the goals and tasks of the two user types. Each user needs at least three tasks (pick one from the provided ones above and come up with at least two more on your own).
  Subsequently, reframe the users' tasks from domain-specific tasks into abstract form. Check out Chapter 3 in Munzner's book, especially the section on "why" a task is performed. Each task should have a number, and your report should connect each view to one or multiple tasks. (Alternatively, you are welcome to also approach it from the tasks of Heer+Shneiderman, 2012)
• Design two dashboards, each one supporting the tasks of the respective user group, with multiple linked visualization components, and describe the interactions between these individual components. Try to find different ways to solve the tasks – the idea is to explore the space of potentially practical solutions with the help of these two dashboard designs. Please avoid basing your second dashboard design on the first one (e.g., design one and then an obvious sub-standard one), as this might result in a deduction of points.
  To kickstart your creativity, the Five Design Sheet (FdS) Methodology by Roberts et al. could serve as an inspiration: FdS. If you decide to use the FdS Methodology, include only the last page (the realization design) into the maximum 8 pages of the report and add the remaining 4 pages (8 total for both designs) to the document as an appendix.
• For each dashboard, argue why you have chosen a specific visualization for each task. Describe how the visualizations are interconnected.
• Critically reflect on your dashboards: What are the pros and cons of each dashboard? Link them to the design principles introduced in the lectures.
• Conclude with a summary of the most important aspects of your report, what dashboard you chose for the implementation in A5, and your reasons for this decision.

Grading

The grading will be as follows (100 points total):

Data, users, and tasks (25 points)

Please create a detailed description of your users and enrich the information about the 2 distinct users mentioned above and their corresponding tasks (min. 6 tasks total). Give a detailed data description (interesting features, missing values, etc.) and link it to the tasks. To help your users do their jobs, you should verify that your data selection supports the requirements of their tasks. For example, you could load the data into Tableau and inspect the data closely (inspect given features, missing values, etc.).

Task abstraction (15 points)

Reframe the users' tasks from domain-specific language into an abstract form, using action-target descriptions (see Chapter 3 in Munzner's book or task lecture (link); alternatively, you are welcome to also approach it from the tasks of Heer+Shneiderman, 2012). Each task should have a number assigned so you can easily refer to it in the following sections to argue which visualization supports which task.

Functionality and usability of the dashboards (10 points each per dashboard = 40 points)

Design two dashboards, each one supporting the tasks of the respective user group, with multiple linked visualization components. Please describe and illustrate each view as well as the overall dashboards and describe the interactions between their individual components. For the illustrations, you can use a drawing tool or paper & pencil, which has the advantage that you are only limited by your imagination as to what is possible. Each dashboard must contain at least 4 views, with at least 3 being different visualization types.
For each dashboard, argue why you have chosen a specific visualization for each task at hand (use the numbers from Section 2). Be sure to describe how you will use interactivity and linking between the views. If there is no linking between the views within a dashboard there will be points deducted. Each dashboard will be evaluated based on (a) functionality (how are the tasks implemented in the dashboard?) and (b) usability (how usable is the design with respect to marks & channels, and effectiveness & expressiveness?).

Reflection (15 points)

Argue about the pros and cons of each dashboard and the respective visualizations. Here, we are looking for good argumentation about the various designs' strengths and weaknesses concerning the tasks but also concerning the design principles discussed in class:

• What do users get done efficiently and fast with your design?
• How did you balance the most effective visual encodings with the number of attributes in the data?
• Which views are overview views, and which are detail views?
• What tradeoffs in interactions did you make? What limitations are there?
• How does each dashboard link its views?

Conclusion (5 points)

Please argue and summarize (based on the design principles and tradeoffs discussed in class and on your reflection above) the most important aspects of your report.

Failure to observe the submission instructions will result in a grade of 0%.

Common Pitfalls

Please make sure that you avoid the following pitfalls. We will take points off for

• making the user scroll up/down
• making the user switch between tabs
• use of drop-down menus (without strong justification)
• [un]readable labels
• using "_" in labels
• missing axis labels
• human [un]readable numbers
• if NULL/missing values are shown as zero
• missing interactivity between charts
• animation that is not well justified (and which alternatives were considered and rejected for what reasons)
• using popup views (without strong justification)

Submission instructions

Submissions should be done via Moodle

Please use this template and submit it as a pdf file named “a12345678_A4.pdf” - replace a12345678 with your student number - to Moodle.

Late submission

Late Submissions are possible, you have a total of five grace days for all assignments. After these days are used up, remaining assignments must be submitted on time.

Academic Honesty