Making a slide deck

In a little over two weeks, I’ll be presenting work with two colleagues on re-designing our introductory course to focus more on data science, whatever that is [1]. To help enhance the quality of the presentation, I decided to give a draft presentation at today’s CS Extra [2]. Of course, that meant that I had to write the presentation.

My preferred mode of presentation is either whiteboard or live typing. In fact, back when I was on the market, I gave a job talk as a chalk talk rather than a set of transparencies [3] because the institution did not plan to have me teach a class and said, we’ll learn how you teach by watching your presentation. Since I did not (and do not) teach with prepared slides, I thought it was best to talk live. It’s harder to do that once you’ve prepared your transparencies (which I had done for other schools), but they thought I wasn’t taking it seriously.

But slide decks are the normal mode of presentation at my professional conferences. And a good slide deck can make a good presentation. But you know what’s not a good presentation? A bunch of bullet points that you read. As I learned from Scott McCloud [4], slide decks are better when they focus on images, rather than words.

But images are hard to put together. I can write a five-bullet-point slide in under five minutes. Making the equivalent image slides usually takes me closer to an hour. Why? I need to decide on the right image, find it, cite it [5], possibly edit it, think about additional images that should share the slide, and so on and so forth. when things come together, it’s worth the effort.

In any case, I started preparing the draft talk at 9:30 a.m. I finished at about 2:00 p.m. That feels like a lot of work for a twenty-minute talk. And, unlike time spent musing, I didn’t find that time spent preparing the talk gave me any further insight into the work. When I write, I think about content. When I design presentations, I think mostly about presentation and organization.

It seems that the talk went pretty well. I made a few jokes that will only make sense to the Grinnell audience [6], but I don’t plan to use them in the main talk. I spoke a bit fast. I learned that one of my students now uses the Racket Plot package for their Econ homework.

And, not so surprisingly, I realized that almost every time I decided to use bullet points rather than images, I should have used images. So now I have to go back and convert two or three slides of bullet points to ten or fifteen slides of images. At least I have two weeks.

Then I’ll give the presentation and, um, the deck that I’ve worked so hard on will no longer serve a purpose.

What’s the point again?

Oh, I remember. There may even be a few. A good slide deck might engage my audience and get them to think more about our approach [7]. Maybe it will even get some to read the paper and, if we’re lucky, they’ll decide to try some of the ideas. I’ve also created something and should take pleasure in the experience of creation.

You know what? If I can choose my creative activity, I’d much rather muse or make a sculpture.

[1] Yes, data science, whatever that is is a theme of the presentation.

[2] I had hoped to do it next week. But some conflicting engagements moved it to this week.

[3] Transparencies are somewhat like Powerpoint decks. Except you put them on a projector. And you often generate them by hand. I think Powerpoint was designed to mimic them. Transparencies have the advantage that you can write on them live.

[4] Or, more precisely, as I learned at a talk by Scott McCloud.

[5] Yes, I do cite the images I put in my slide decks. It’s the responsible thing to do. I hope that you do the same.

[6] Such as a reference to Hawkeye vodka.

[7] Or at least the issues I’d like them to consider.

Version 1.0 of 2018-02-08.