Module Three: Design Workflow

due by 11:59pm on April 9 (note that this will be our final assignment before the Final Project)

submission under "mod3writeup.html"

Design studies are intensive, longitudinal, collaborative design activities. While they're a powerful tool in your visualization development toolkit, they also take a bit more energy than is reasonable for a module-sized assignment :-) To get you some practice with the design study methodology, you'll conduct a miniature design study with yourself as both the designer and stakeholder (with some help from external sources). You will submit your design study as a PDF write-up with each of the numbered components below (note that it's totally ok to keep your responses brief). Please feel free to add any additional materials to your submission repository as usual and be sure to include any relevant materials like source code for running your visualizations if you choose to include interaction or other components in your design.

The Problem:

Design studies and other visualization design workflows are inherently task-oriented: we want to understand what it is people need to do with their data and build tools that let them achieve those goals. However, true design studies often take years to complete. In this interest of brevity, we're going to tackle a challenge where at least part of the work has already been done for us. You can use one of two datasets:

• The data you used for Assignment 2
• A dataset discussed in an article on 538

The reason we're using 538 is simple: they provide open access to the data underlying their articles. You can link out to the datasets from most articles or use their web (Links to an external site.) or GitHub (Links to an external site.) repos. You can use other sources, but check with me before moving forward so that I can help you scope things out in a meaningful way.

You will choose a dataset and try to design a tool to probe the data in more detail. In other words, your goal here is to build a visualization/set of visualizations that would allow someone to explore that data in more detail to either come to their own conclusions about the data and problem described in the article (538) or that you've started to construct (Assignment 2). If you use your Assignment 2 data, you've got a good jump already on what the interesting bits might be. If you use 538, you can use the write-ups in the accompanying article (plus outside sources if you'd like) to deconstruct what the core questions are in the data.

You will complete the following steps to design this system:

Precondition: In a typical pre-condition phase, you'd be both reading about your target subject and talking with people. Since either you or the expert at 538 has already taken care of that, we'll skip the casting phase and instead focus on learning about your problem area and winnowing down to a target dataset.

1. Learning: In your write-up, provide a brief background for your target problem. Make sure to describe the problem and why it matters.

2. Winnowing: Describe the data that you will use to explore that problem. Make sure to talk about the background of the data (e.g., where does it come from, what bias might it have, etc) and why it's useful for your question as well as including a link to the source article.

Core: Once you've decided on the dataset, you'll want to assume the role of the designer. You will work through the discover, design, implement, and deploy phases to create a visualization that addresses your target problem. You will achieve this by doing a mini-iteration on the following steps:

3. Discover: Characterize your problem by identifying at least two tasks that you will want to conduct with your data, drawing either on your prior knowledge or the content of the article to scaffold your investigation. Enumerate the tasks using the why/how/what/where/when/who framework we talked about in class (Lecture 15 has more details).

4. Design: Create a sketch of your initial prototype. Note that this sketch can be drawn from your work either in Assignment 2 or in the source article, but should include some design iteration to support your target tasks. Add brief design justifications and discussions of the trade-offs of key design choices in the sketch, being sure the discussion is closely tied to your target tasks. Make sure to include a copy of the sketch in your write-up.

5. Implement: Implement your design as a more polished digital or physical data representation. You can use any tool of your choosing, but if you elect to use a WYSIWYG tool like Tableau or Excel, please incorporate design choices beyond the defaults provided by the tool. Note how you've implemented your solution, how it addresses your core tasks, and include at least one image of the visualization in the write-up.

6. Deploy: Use your visualization with your target data to conduct the tasks you outlined in Step 3. Note your observations about the data gained through these tasks.

7. Iterate: Note at least one new task you'd conduct now that you've had a chance to investigate your dataset. Describe how you would change your solution to accommodate that task.

Analysis: Once you've completed your core phase, you can reflect on designing your tool. We'll skip the write phase since you're already doing it in creating your write-up!

8. Reflect, Pt 1: Describe what your solution tells you about your target problem. Note that you will do this with your domain expert hat on.

9. Reflect, Pt2: Describe what your solution tells you about designing visualizations for your target problem. Note that you will do this with your designer hat on.

Rubric:

The rubric for this assignment is as follows:

• 1 pt for submitting a written document and any accompanying materials (e.g., source code)
• 1 pt for each of the nine write-up points above (9pts total)
• 10 pts total

Bells & Whistles:

Live Deployment (1pt): Deploy your design study as part of a live web portfolio. Provide a link to that portfolio somewhere in your submission document.

View Coordination (2pts): Your tasks may require multiple perspectives on the data. Include a visualization that has at least two views that collectively address the target task. Design your visualizations to interact with one another using view coordination. That is, interacting with one visualization will cause a change in another.

Iteration (2pts): As part of Step #7, restart your design study to redesign, reimplement, and redeploy a visualization updated to support the new task. Make sure to include both the original discussion, describe your reimplementation, and add a new response for Steps 4-6.

Above and Beyond (3pts): Nothing extra to do here: these are points awarded at the graders' discretion for truly outstanding work.