(The title of this post is a play on the title of a famous book that I have never read)

CS765 (Data Visualization) will be taught in an online format this fall. This posting will tell you why I have chosen to do it this way. It will not tell you how the class will run, because I haven’t completely figured it out yet. But understanding why, will help you understand what the implications are for the class.

A big part of this is me thinking out loud to convince myself I am doing the right thing, but it should also give you some ideas about how I think about the class.

Why choose online?

CS765 is scheduled to be small enough that I could have held it in person. I chose to hold it online.

The biggest reason: Safe Interactions Between Students.
A lot of the class is about interactions with other students. This would be hard in a “socially distanced” classroom. Normally, CS765 is held in a “collaboration classroom” where students sit around tables so they can interact with each other. This kind of interaction seems infeasible - we can do a lecture in a big hall, but having people talk to each other? From a practical standpoint, having lots of small groups speaking to each other only works when they are close enough that they can talk quietly. From a COVID safety standpoint, I don’t want to force people to have close interactions (in-person conversations, collabroatively using paper, …) with people they do not know.

In theory, online tools should be great for these kinds of collaborative activities.

• We already use online discussion tools for written discussions.
• Video conferencing tools allow people to be put into small groups for discussions.
• It is easier to randomize / shuffle groups, so you will get more variety in the people you talk to.
• It is easier to log the work you do
• Shared editing tools allow for people to work on things simultaneously, add comments on others work, …

There will be kinks to work out. It’s all an experiment. For example, I have no idea how we’ll do the collaborative design exercises that have been a big part of the class. But these wouldn’t work in person under COVID anyway, since they involve sharing paper and pencils.

I think this may actually lead to a better class (modulo the experimental nature of figuring out how to execute the class):

• It should take the emphasis away from lectures, and put it more on “activities” where you get to learn by doing.
• It should help prevent me from using too much of class time for lectures (it was very easy for me to ramble on ad infinitum - I love to talk about vis stuff, so once I get started…)
• It lets us be more creative in how we do interaction
• It should let us make class time more interactive, not less

For me, moving the class online is extra work. I’ve run this class enough that I know how to make it work in person. In person lectures are easy for me. I have basic designs for in-person collaborative assignments. I know how to use the classroom technology (the collaboration room has some interesting tech). But, I think the practical issues, and the potential for a better class, will be worth it.

Some other reasons for choosing to be online (in no particular order):

• While a spread out classroom is (in theory) safe, it still requires people to act responsibly.
• Part of the argument for live lectures is that the lecturer can “read the audience” - but I’m not convinced that’s possible if everyone is wearing masks.
• We would have to shift to online after Thanksgiving, planning to have class both ways seems like extra hassle.
• One of the worst aspects of online teaching in Spring 2020 was that I had to switch to it without much warning. By starting online, I am ready in case something happens where the University needs to shift to online in a hurry.
• I think that once I figure it out, it could lead to a better class (and those ideas can be used in the future)
• Even if there was an in person component, its not clear that I can “require” it - which means I need to provide online options anyway. And consistency is good for many reasons.
• I am not sure how I am going to feel about going to campus this fall. I might prefer to work from the safety of my own home.
• In theory, students can take the class even if they are not coming to Madison.

How will the class work?

I am not sure how the class will work as I type this (July 25, 2020). I am putting down some thoughts as part of my efforts to organize it.

Some basics:

• There will be a synchronous component to class. Expect to “attend” online during the class time (Wednesday/Friday).
• There will be online assignments
• There will be a sequence of smaller projects
• There will be a substantial amount of reading (this class is notoriously a lot of reading for a CS class - when humanities and social science students have taken the class they laughed at that)
• There will not be a synchronous exam. We might have some for of “summative activity” at the end, but it will not be a scheduled exam in an exam time slot.

But, before I can say more, I need to explain the class a bit.

How did the class work in the past?

Before considering how the class runs online, you need to know how the class ran in the past. You may want to see last year’s course web for details.

This class is different from most CS classes in that the emphasis is not on the technical components. (although, there are plenty of those)

A lot of the class is about getting students to think about visualization, to critique it, and to consider how to design it. Rather than big programming projects, there are a lot more small exercises that try to re-enforce skills like design and critique and seeing how the ideas from class are applied in the world. See the 2019 What is this class and Why posting. This years What Is This Class and Why? is nearly identical.

Last year, the class met twice a week (although, it was “over-scheduled” for 3 times a week)

• One 75 minute period (usually Monday) was a “lecture”
• One 75 minute period (usually Wednesday) was an “in class exercise” (ICE)
• Some weeks, there was an optional third class that talked about a topic related to an assignment

In addition to that, each week there was:

• Required readings
• A required online discussion
• A required “seek and find” (an online assignment)

And, there were:

• Three “Design Challenges” (mini-projects), although one was very different than the others, and in prior years there was a fourth even more mini assignment

These seven parts of the class for the in person class were explained on the 2019 Parts of Class page. (and are explained on this years Parts of Class). I expect there will be a similar page this year (once I figure out how I going to do things). The important thing is that all 7 of these pieces are important to the class, and I plan to adapt each one to the online setting.

How does that translate to Online?

Before translating this to online, it’s worth asking should this design be translated to online. I think the class is successful for many students. I see students coming out of the class having achieved my goals (both from talking to students, as well as from working with them after class). But, that doesn’t mean things couldn’t be better.

A lot of the concepts and skills in class are best learned by interacting with others - doing collaborative design activities, performing critique, discussing things to help shape your viewpoint and check your understanding, … The in-person design of the class tried to do this through online discussions and in-class activies.

I am skeptical of the technology for online discussions (live, audio/video). However, it doesn’t work perfectly in person either. Students don’t always like to interact with a lecturer (and there are timing issues, and…). And for discussions among students, when you have 50 people trying to talk to each other it becomes quite a cacophony pretty quickly.

So, some of the things that could be different:

• The lectures are pretty non-interactive. They were often monologues. They were dumps of information. I am not sure that’s the best way to get it across to people.
• I had no way to force people to learn the topics from the readings. The lecture/reading topics are indirectly tied to what we really do in class. (I tried quizzes, but that was a lot of work for everyone, and not hugely successful)
• We didn’t really assess and evaluate the in-class assignments. We didn’t return to reflect on them (so the lessons really come out), and the ungraded nature meant that people didn’t take them as seriously.
• Online discussions (asynchronous) weren’t a very successful way to get students to interact productively around the topic.

Thinking about the components of class and how to make them better suggests:

• Lectures: may be better asynchronous (pre-recorded). That’s more work for me, but they aren’t that interactive, and we can have the interaction as a separate piece. With a pre-recorded lecture, you can always pause and take a break, watch at 1.5x speed if I am too boring, or review to catch the things you miss. And then we can interact around the material in other ways (explicit discussions, …)
• In class assignments: these need to be adapted to small groups online. This will be tricky, but it has the advantage that we will be able to record more of what happens (e.g., the designs that students create).
• Optional tutorials: these can still work online. It’s not as good, but from experience with CS559 online, they can be made to work. They are like lectures, but are places where even limited interaction adds a lot.
• Readings: these have always been online - we’ll have some component to see that you’ve actually done the readings
• Discussions: I am thinking that these can be moved to be more into synchronous, live discussions. But there will be at least some online discussion
• Seek and Finds: these worked well, and won’t change much (other than I might try to integrate the results into class more)
• Design Challenges: One aspect missing previously is the class was presentation / discussion of student work. With online, we may actually be better able to do this because we can divide the class into smaller groups.

So, I am optimistic that we can make use of the upsides of online (ability to have some components be asynchronous, ability to split people into smaller groups for synchronous discussion, ability to interact “multi-modally” online, …) to make the class work well.

Why the odd timing?

The class meetings are scheduled for Wednesday/Friday 11-12:15, which is a pretty non-standard time.

Initially, this was an artifact of scheduling: when I scheduled CS765 originally (when everything was still “normal”), the only way that I could get the room I wanted (one of the “collaboration classrooms” in Wendt library) was to have this non-standard time. Since 11am is my favorite time to teach, I decided to go with it, despite the odd days.

When we switched online, I decided to keep the same time slot - mainly for simplicity. I wasn’t sure how I was going to use the time slot (since I wasn’t sure if the class was going to be synchronous or asynchronous), and it seemed simplest to avoid changing things - everything else in the world was changing, it would be nice to have something hold constant.

However, as I think about it, this might actually be a good way to schedule the class: we’ll use the beginning of the week for asynchronous activities (reading and video lectures), so later in the week you are prepared to discuss things.

Gathering of stuff for students

• First, online students have more autonomy and responsibility for carrying out the learning process.
• Second, online students must be able to use technology to access resources and communicate with others.
• Third, they need the self-discipline and motivation to persist in learning when there is less direct contact with an instructor or other students.