Arrivals Exercise

in Posts
Page content

As an experiment, we will do an “in-class design exercise” as an online discussion (not in lecture time). I described this in class on Friday, October 2nd (see (05-2-Critique-Arrivals.pdf 0.7mb), or you can review the video recording of the lecture).

The goal of this assignment is to get you to integrate thinking of scale and comparison into your visualization design process, and to give you a concrete example to consider the frameworks for comparison and scale. It also is meant to give you practice with a design problem, including discussion and critique around tasks and designs. Design challenge 2 will resemble this exercise.

The specific problem considers lists of arrivals. Two model problems are “when do students arrive in class” and “when are assignments handed in”.

In the past, this has been a very successful in-class exercise. Trying to fit it into the cadence of our class (with postings due on Tuesday and Friday, lectures on Wednesday and Friday) is tricky because we have limited time for iterative cycles (which are hard to do asynchronously). We’ll see how well this works.

You must draw your sketches (as you would have in class, no programming!) These should be “quick” sketches, don’t spend hours making masterpiece artworks. Do what it takes to convey your ideas to others.

Part 1

Consider the simplest variant of the problem: you have a list of arrival times for a single event. You can assume there is a “start time” (e.g., the official start time of class, or the due date for the assignment), so each arrival might be early or late.

In this part, consider a single event and 30-50 arrivals. (e.g., the students arriving at class on a particular day). The data is a list of times.

We discussed some tasks in class.

Part 1A - Sketch 2-3 designs for this data, considering some of the tasks. Give a short (a sentence or two) rationale for why the design addresses some task(s).

Part 2

Consider the variant of the problem where we have an ordered list of events. For example, in our class we have had 10 lectures, each one had about 50 arrivals.

Note: you have limited data. You don’t know anything about the arrivals. For example, you can’t tell if the same person comes early every days (since you know a time of arrival, not who arrived). We’ll relax this later.

For this part, think about scales on the order of our class so far: a few (3-10) events, a few dozen (30-50) arrivals per event.

Part 2A - What tasks might you consider with this data?

Part 2B - Sketch some designs for this data - Each student should sketch out at least 2-3 designs, and give a short (a sentence or two) rationale of why the design addresses some task(s).

Part 3

This problem scales up in several ways (this isn’t even a complete list):

  1. We can have more events
  2. We can have more arrivals per event
  3. We can have longer durations (the range of times that arrivals can occur for an event)
  4. We can have more data about each arrival or event

Let’s focus on #1 (more events) and #2 (more arrivals per event).

For example, consider:

  1. More events (our class at the end of the semester (28 events), or the entire 10 year history of classes (280 events))
  2. Bigger classes (a CS undergrad class with 200-500 students, a concert with 1000-2000 people, …)

You can also consider scalability types 3 and 4 - but you must consider types 1 and 2.

Part 3A - consider the tasks and designs from part 2: how well do they scale? Do the tasks change as the scale gets larger (in either dimension?) How well do the designs scale as the data grows in the various dimensions?

Part 3B - develop some new designs for the “larger” data. You should come up with designs that scale to hundreds of arrivals, dozens of events, or both at the same time. For each design, describe the tasks that your design is meant to address, how the scalability strategies inform your designs, and some rationale for why you think your designs will address the tasks at scale. You should provide 2-3 designs. If you think your designs from part 2b scale well, you can re-use them (but be sure to provide the rationale).

The Plan

With each phase, there is the opportunity for discussion and critique. Each phase builds on the previous phase, so they need to be spaced out. I am also aware there is a Design Challenge due, and have adjusted accordingly.

Roughly, Parts 1 and 2 will be the online discussion for Week 6, and phase 3 will be the online discussion (or part of it) for Week 7.

Week 6 (October 5-9): We will use Online Discussion 06: Scale (due Tue, Oct 6) for Parts 1 and 2.

Over the course of the week, you will need to make 3 “initial” postings (postings not in response to other postings):

  1. Part 1A - Sketches of 2-3 designs, with rationales. You should use your preferred method for Drawing (Art Supplies) and method for Scanning to create the designs.

  2. Part 2A - A list of tasks for this data type.

  3. Part 2B - Sketches of 2-3 designs, with rationales.

1 and 2 correspond are the Tuesday requirement, while 3 is the Friday requirement. We will be lenient with the Tuesday deadline - since we know there was a big deadline Monday, but we do want you to complete parts 1 and 2 promptly (say, Wednesday), so you can discuss them and refer to them for Part 2B.

This is intended for discussion: please discuss tasks, working together as a group to develop a task list (which you will need for 2B) and critiquing each others designs (although, we often see many similar designs).

Week 7 (October 12-16): We will use Online Discussion 07: High-Dimensional Data (due Tue, Oct 13) for Part 3. We will connect common strategies seen in designs to the strategies in class.

Some Pedagogical Notes

This exercise is designed to fit with the class material on comparison and scale.

In the past, I provided the comparison framework and ways to thinking about scalability in class first, and the exercise was an opportunity to apply those concepts.

This year, the exercise comes first: the hope is it will give you a concrete problem that will help you learn and appreciate the abstract frameworks. That said, you might want to do the readings on comparison and summarization strategies ( Readings 06: Scale) to help you think about how to articulate the tasks and choose design strategies.