Critique: Community influence

Today’s poster was sent in by kindly contributor David Selby. It was created for the useR! conference in Brussels earlier this year. Click to enlarge!

The main data visualization gives this poster a strong graphic element at its core. The visulizations almost look like abstract art. David did the right thing by making these as big as possible. You wouldn’t be able to interpret these otherwise.

David has skillfully mixed both a serif and sans serif font in the type in a way that is not distracting.

There may be a mild problem with reading order. Looking at the text, this was the pattern I expected to follow:

Instead, I realized that I was supposed to go like this:

In fairness, the acknowledgements can be skipped, so I don’t have to drag my eyes all the way back to the lower left. But still, I was confused when I realized that black of text was acknowledgement. “Wait, I’m not supposed to read this yet!”

David was very clever to link the “Web of Science” data and “Statistics” data using colour. But it still bothers me that the two “Statistics” graphs are spatially separate, rather than adjacent.

David has a brief blog post about the poster, and wrote:

One of the key things when doing the analysis was to keep everything reproducible. To this end, all code for the graphs and results is presented in a GitHub repository and vignette, along with the Scribus file for the poster itself. All software used was free and open source. Modulo the raw data, anybody can recreate the design and repeat the analysis for themselves. I also used the vignette to track my ideas during the design process and list some sources of inspiration, even though it’s not really relevant to the actual research.

Here is the poster on the board (photo: Oscar de León):

I am pleased to report that this was an award winning poster: first place (shared with two others, like the Nobels)!

External links

useR! poster: ranking influential communities

Comments

Scott McCloud’s “Big triangle” and poster design

Posters are a visual medium. But not everything is equally visual. A picture of a real object is very visual, and the best thing to have on a poster. A scatter plot is less visual. And text is the least of all. I was thinking about how I might make that point, um, visually, and I suddenly realized that I was just recreating one side of Scott McCloud ’s triangle from Understanding Comics . If you have not read Understanding Comics ... oh, how I envy you. You have that to look forward to. It is a wonderful book. Even if you are the sort who thinks, “Ugh, superheroes,” get over it, read this damn book, and have your consciousness expanded. It is an undisputed classic book. Here’s a except relevant to the matter at hand: And that’s the point I was trying to make, except McCloud did it better over twenty years ago. Received information is immediate; perceived information takes effort. This is why nobody likes posters with too much writing. It takes effort that, in a busy conference settin...

Link round-up for March, 2019

Mike Morrison has a 20 minute video describing what a poster session is and how to make a poster. Unlike Mike, however, I do not believe poster sessions are “holding the human race back in a non insignificant way.” The video, particularly the first half, is pessimistic about poster sessions. Around 13:25, good stuff starts to happen as Mike outlines a good poster design. I think he overestimates people’s willingness to snap pictures of QR codes, though. Mike has provided templates here , and is working on a study validating the design he has. He is looking for grad students to participate. You can email him at Mike.A.Morrison@gmail.com . Amy Burgain saw this video and offers this alternative : Amy writes: It achieves the clear simple message BUT emphasizes how that message is supported by the DATA. It keeps the goal of understanding how the conclusion is related to the data while also making it easier to glean main messages. I plan to have my own longer post about this in a couple of...

You have options for numbers (PowerPoint users need not apply)

If you must have a table on your poster, look into what options you have for your numbers. Many fonts have number variants. Proportional numbers have skinny numbers (e.g., 1) and wide numbers (e.g., 0). Two numbers differ in width depending on what numbers they have. But tabular numbers are all the same width. So decimal places and dividers will line up if the numbers are lined up, as they are in a table. If you have a table, it only makes sense to use tabular numbers if you can. They are explicitly designed to make your tables more readable! But tabular numbers will only do so if you follow a couple of other good practices: Make your numbers right aligned. Use the same number of decimal places in each column. You may also find a couple of other options. numbers can be either lining numbers (all the same height) or oldstyle (with ascenders and descenders, like upper and lower-case letters). That means you have four options for many fonts. In Microsoft Office, these options are...

publisher tips

Search This Blog