Initial Posting Due: Tue, Sep 19 at (Canvas Link)
Readings
The topic for this week’s readings is Abstraction – especially data abstraction.
- Shneiderman, B. (1996). The eyes have it: a task by data type taxonomy for information visualizations. In Proceedings 1996 IEEE Symposium on Visual Languages (pp. 336–343). (doi) (web pdf)
This is a classic. Possibly one of the most influencial papers in the field. It’s old, and newer things are far more extensive. And the field has moved on from 1996 in many ways. But the initial thinking of abstracting data and task separately, and suggesting what those abstractions might be, really started here. The information seeking mantra is a classic notion. This paper is dated enough that it can be hard to read – but it is short.
-
What: Data Abstraction (Chapter 2 from Munzner’s Visualization Analysis and Design) (Munzner-02-DataAbstraction.pdf 1.1 mb)
A fairly dry description of the types of data. Don’t worry about trying to remember all the terms – you can always look them up when you encounter them again.
Despite it’s length, the chapter skips a key concept: level of measurement for scales. You might have learned this in a stats class, but please understand the difference between “scale types” (nominal, ordinal, interval, ratio). Usable Stats has a simple introduction.
-
Why: Task Abstraction (Chapter 3 from Munzner’s Visualization Analysis and Design) (Munzner-03-TaskAbstraction.pdf 441 kb)
Figuring out how to think about tasks is important. This chapter (and the research paper it is derived from) focuses too much on trying to put every task in a neat organization. What’s important is to think about tasks. This is one way to do it, and it will help you learn to think about tasks. Don’t get too bogged down in all of her categories.
We’re reading the book chapter, not the paper. I recommend the Schulz et. al paper below for contrast.
-
Forms and Functions (Chapter 2 of The Functional Art) (theFunctionalArtCh2.pdf 8.2 mb)
Cairo’s thinking about “the shape of data” is another way to think about data abstraction.
-
Mackinlay, J., Hanrahan, P., & Stolte, C. (2007). Show me: automatic presentation for visual analysis. IEEE Transactions on Visualization and Computer Graphics, 13(6), 1137–44. DOI (File on Canvas)
This is a “modern” research paper, but it’s an unusual one. It’s easy to dismiss this paper as marketing for Tableau – but it really does give a sense of how good abstractions can help in choosing appropriate visualizations. Plus, we’ll probably use Tableau this semester, so learning about it is a good idea.
Because Tableau is such a direct implementation of the “building blocks” theory of visualization, it provides a great way to experiment with it.
While it isn’t technically “reading,” part of the assignment for this week is to start looking at different kinds of visualizations (especially standard chart types) and trying to understand what data types and tasks they are good for. We’ll continue this next week when we connect these different visualization types to the visal pieces they are made up from.
Here are a few places to look for catalogs of visualization types:
- The Data Visualization Project – a nicely done visual catalog of known visualization types. I like this a lot. A mix of obscure and standard designs. The discussions are useful.
-
A tour through the Visualization Zoo. By Heer, Bostock, and Ogievetsky. There are various versions around. But, the one in ACM Queue online (here) is good enough. This is a more academic paper, that has various forms.
-
The D3 examples page has hundreds of examples of things – some standard, some not-so-standard.
-
The visualization taxonomy developed for the MassVis project groups many standard designs. It tries to organize the different designs.
-
The periodic table of visualization methods and the Data Visualization Catalogue also have lots of examples of different types of visualizations.
Optional
That’s already a lot, but understanding task is really key to doing visualization well. These papers are strongly recommended.
- Schulz, H.-J., Nocke, T., Heitzler, M., & Schumann, H. (2013). A Design Space of Visualization Tasks. IEEE Transactions on Visualization and Computer Graphics, 19(12), 2366–2375. (doi) (web pdf)
This paper came out at the same time as the paper behind the book chapter. It was literally in the same session of the conference. I actually find this to be a more useful way to think about task – it’s not as encyclopedic, but that’s a feature.
-
Sarikaya, A. and Gleicher, M. Scatterplots: Tasks, Data, and Designs. IEEE Transactions on Visualization and Computer Graphics, 24(1) — Jan 2018 . (web page)
An upcoming paper that my student and I wrote. It focuses on a specific (but ubiquitous) kind of visualization, but thinks through the tasks and shows how thinking about the data properties and tasks helps suggest designs. I like this paper, but I am biased.
Online Discussion
Initial Posting Due: Tue, Sep 19 at (Canvas Link)
The topic of this week is abstraction. This is a central and important topic, but one we often take for granted. It’s not as glamorous a topic to think about (as say perception, or specific design types), but it provides a critical foundation.
The idea of this discussion is to make sure that students understand the two key types of abstraction (Data and Task) that are critical for visualization. If you’re a computer scientist or mathemetician, you are probably pretty fluent with the concept of abstraction – even if you don’t think about it explicitly.
- Data abstraction is key because it lets us map our visualization designs to the right kinds of data. When there are mismatches, there are problems.
- Task abstraction is key because it lets us see how general solutions can map to many specific problems.
Thinking about data abstractly is easy (or seems so to me). Thinking about task abstractly is more challenging, and it’s only in the past few years has the visualization community come up with good ways to talk about it.
There are many ways to think about tasks abstractly. I haven’t seen one yet that totally nails it. Munzner’s (which actually comes from a longer paper where they have an even more complete model) is about as good as I’ve seen so far. But view it as a structure for thinking about task, not the definitive way to do it.
So this discussion assignment has the twin goals of making sure you think about data abstraction and making sure you think about task abstraction. I’d like you to try to do this for 2 different visualizations. (You’ll also do one for the seek and find)
For your intitial postings, I want you to pick a visualization (in the style of a seek and find – please either upload a picture or give a link) and:
- Describe the DATA abstractly
- Describe some TASKS concretely
- Describe these tasks more abstractly (in your own words)
- See how these tasks fit into Munzner’s taxonomy (or not), or one of the other taxonomies
Since you have to make 2 initial postings, you’ll need to do this twice (so this week you’ll do at least 3 since there’s the seek and find). If you want more practice, feel free to do more than two.
For discussion, comment on other people’s abstractions – do you agree? Can you categorize the data more specifically? Can you identify alternate possible tasks? Can you think of different ways to abstract the task? Which abstractions do you think may be useful in helping to choose solutions?
One tricky thing: when we see a visualization, we don’t know what the designer was intending for us to do with it – so we don’t necessarily know the task it was designed for. So, in an exercise like this we are either (1) looking at the tasks that are facilitated by the visualization or (2) thinking of tasks we’d like to do with the data/visualization (but may not be able to). Either of these is OK – we’re not always saying that a visualization succeeds at enabling the task.
As an example… consider the example from the first week in class (also described in the Simple Example: 4 Design Moves posting) in looking at the rounding errors in grades:
- The data are records (it’s a table) corresponding to students, although I am really only looking at two values per student: computed grade and assigned grade. Both of these are quantitative values. I think of them as interval, rather than ratio scales (i.e., it’s hard to say an A (4.0) is twice as good as a C (2.0) – it’s like temperatures). One is continuous, the other is descrete.
- The task I described was identifying students who were hurt by the rounding errors when we assigned the quantized grades.
- A more abstract description of this task is to identify/examine boundary cases in grouped data.
In Munzner’s taxonomy, this might be an “Identify” Query task. (although, there are some other categories you might argue it falls into).
