30 May 2019

Link roundup for May 2019

In our semi-regular “Best poster I’ve seen” dredged from social media, Fabian Roger calls this “absolutely brilliant”:


Poster by Therese Karlsson, who said, “Never had as much fun making a poster as this one!”

A longer post analyzing this poster is coming!

• • • • •

Mike Morrison’s billboard format poster continues to generate discussion. My Cousin Amygdala wrote:

This whole Instagram-a-fication of scientific posters movement is my least favorite science trend right now. It misses the entire point of scientific posters, which is to facilitate networking and scientific discussion. Your poster isn’t a popularity contest or a memory test. You present a poster at a meeting to facilitate discussion with your peers. It isn’t a passive process. No one is going to “discover” your poster.

You are presenting at a scientific meeting, be proactive about using your poster as a networking tool. Tell interested parties about it beforehand, nicely, and that you would appreciate their input. If you are a student your mentors should assist you with this process.

I have attended countless poster sessions and thrown away enough poster handouts to cover the earth three times over.

I do not remember the contents of any single poster, but I absolutely remember & have benefited from the peer interactions that posters facilitate.

The conversation continues in retweet. Hat tip to Dr. Becca.

Here is a short thread from Kathryn Vaillancourt. Excerpt:

The #betterposter seems like an overstep in the right direction.

There is also lots to consider in Julie Blommeart’s thread. Hat tip to Milton Tan.

• • • • •

Speaking of Mike Morrison, he is interviewed about conference posters here. Excerpt:

I’m very personally invested in improving the rate of discovery across many different fields. As for the poster, it’s something that I could take a swing at on my own. There are much more damaging bottlenecks, like the scientific article publishing system, peer review inefficiencies, etc. But those will take teams of well-funded people to fix. The poster was the lowest-hanging fruit.

• • • • •

Meanwhile, in the, “Don’t be that person who makes graphs like this” department:


This... seems like a lot of work to go through for two numbers. (I duplicated this in Excel, and it took a long frickin’ time.) Hat tip to Louisa Smith and Justin Kiggins.

• • • • •

I have been to India. The women are not as tiny as this graph makes them appear.


I have no such first hand knowledge of women in Latvia, however. Hat tip to Bill the Lizard.

• • • • •

Zombies are hot in popular culture, but should not be in research conferences. Nature has an article about how to avoid becoming a “conference zombie” ( person who has no energy and does not notice what is happening around them).

Plan. Take breaks. Sleep.

• • • • •

Nice article about using colour in data visualizations. Tips include colour coding for bad news:

For negative results red, orange, purple and generally darker and muted colors often feature in data visualizations.

Hat tip to Nancy Duarte.

• • • • •

Speaking of colour, if you organize or recognize books by colour, you are not alone. Helen Rosner wrote:

When I worked at a bookstore, the number of customers saying, “I don’t remember the title but it had a purple spine” points to color actually being a pretty good organizing principle. What my books look like is absolutely part of how I mentally catalog them. At home, I tend to organize by category and subcategory and then by size and color.

(V)irtually all books spend more time being looked at than being read. And the more books you have, the more likely it is. Spare me the wrath of the book snobs who believe any whisper of aesthetics undermines virtue.

I think this is a great reminder of how powerful the visual element of something is more than the words.

23 May 2019

Critique and makeover: Snap, the magic dragon

Today’s contributor is Mathilde Mousset. She sent along two posters and said I could use either, but I thought, “Why not both?” Click to enlarge!


Portrait style posters are often tricky beasts, and this one embraces the format well. It uses coloured bars and fine lines to signals that it is read across in rows. It uses a narrow typeface for headings to make best use of limited space. I like the lightweight type for the main text, which makes the bold emphasized words pop even more.

The take home message at the bottom essentially repeats the point the title makes. While repetition can sometimes be useful, I usually find that removing redundant elements reclaims space that I can use to make things larger, and that turns out to be more helpful.

While this poster is typeset and laid out very well, the most visual part of the poster is the methods section in the middle. This is a little less than ideal, because it is a bit below eye level. I’m surprised that a grayscale illustration of a snapdragon is used instead of a full colour image.

Many of those same features are found in the second version of this project.


Mathilde wrote about this second poster:

Because I had little time, the second heavily re-used stuff from the first, except I added some boxes because I felt it needed a stronger structure. Surprisingly, I kind of prefer the output with the boxes (usually boxes in posters make me feel as in a prison).

While I have complained about the overuse of boxes many times, Mathilde’s boxes work. First, they are not black. They are either light yellow, or a darker shade of the background colour (the blue in the bottom box). In either case, they aren’t drawing a lot of attention to themselves. Second, there is a good amount of white space between the boxes and the text, so nothing feels crowded.

The typesetting is perhaps a little less successful on the second poster than the first. Mathilde uses a modified bullet point list, only using a chevron instead of a bullet.


But the point of having a symbol is to be able to scan down a list in a column. So right justifying bullet lists kind of defeats the purpose of a bullet list.

Here’s a quick revamp keeping those bullets aligned.

At the top of the poster, the second level of bullet points also run into minor alignment issues. Again, if you want people to scan down from one symbol to the next, you want to keep the space between those symbols clear of interference.


Neither the first or second level are set with hanging indents. The first level has lines running back to the chevron symbol. The second level is worse, because the second line of text runs back not to the secondary symbol (a dash), but all the way back to the primary chevron symbol.

Here is how I would like to see those paragraphs set up:


Much easier to scan down the list!

None of these changes alters the number of lines of text. The revisions fit in the space, so they don’t require reworking the rest of the section around the changes.

16 May 2019

Critique: Child care

Today’s contribution comes from Mary Bratsch-Hines. Click to enlarge!


Mary writes:

I’ve realized that although I have removed the boxes, it still feels boxy. This is likely because it’s laid out like a manuscript.

The “boxy” style that Mary mentions is not necessarily a bad thing! Paper is rectangular, and lends itself to rectangular layouts. It is true that it can feel a little boring, particularly with repeated use, but it is better a little dull than something that jumps off the rails trying to be different.

If you are every faced with the choice between boredom and confusion, pick boredom!

Mary’s poster is an excellent example of why so often I end up recommending a three column layout to people. It just works. While there is room to improve on this poster (more on which in a moment), nobody would cringe looking at this poster or think it was an accident gone horribly wrong.

What stuck out at me was how little the poster stuck out at me. The overarching sense you get from this from a distance (or at small size) is greyness. And the few patches of colour are all stuck far away from eye level at the bottom of the poster.

A few things I’d try:
  • Put the picture in the right at the top of the column instead of burying it at the bottom (the Cosmo rule).
  • Crop said picture so that it’s column width. The tighter the grid, the close to God.
  • Put the “Results” heading in line with the headings of the other columns instead of under the figure.
  • Make the title and headings heavier and blacker.
  • Make the graph the width of the column.
Here’s a revision with those changes:

    There is more visual contrast and variety in this version of the poster. Even shrunk down, the title and headings are now bringing in some black to break up the greyness. The grey wall of text is broken up vertically instead of all being at the top.

    This is, as always, a quick and dirty revision. If you enlarge the image, you will see the text in the graph’s axes are distorted. To make the bottom center graph column width, I just grabbed is a stretched the whole thing. This is not the correct way to make the graph fit. It should be resized to the correct proportion in the original graphing software.

    The revision above is much better, but the left side still felt like it was fading away into nothing. So I broke up the greyness by colouring the flowchart boxes.


    I picked yellow because the map to the right of it had a little yellow. In fact, the map could stand to do with a little colour, although we’re probably reaching the point of diminishing returns in adding more colour to combat the greyness. A better step would be to do a ruthless edit to reduce the word count.

    If you think the difference between the first and last version is a big improvement (and I’m vain enough to think that it is), remember how easy it was.

    Six changes.

    That’s it. The poster became much visually punchy and attractive with just six changes. Remember that if you ever feel like, “I don’t have time to make this poster better!” You probably do have time: you just have to know enough to make the right ones. And those are not particularly tough things to remember. In this case, those guides are:
    • Line things up.
    • Put graphics at eye level.
    • Make important things bigger and bolder.

    09 May 2019

    More than just visual cheesecake

    I want academics to create objects with high quality graphic design. But I imagine there are people who might admit that a well designed conference poster (or what have you) may be nice and all, but it’s just visual cheesecake. Sure, you get some personal aesthetic pleasure out if it, but that’s about all. If the content is solid, some might say that good graphic design won't help you meet career goals like job offers, tenure, or promotion.

    I personally cannot think of any anecdotes or data supporting the contention that excellent graphics have significantly advanced or hindered individual researcher's careers (e.g., grants or papers rejected just because some aspects of typography or design were so bad). If anyone knows any examples, I would love to hear them!

    On a larger scale, some might argue that the quality of graphic design has neither significantly advances or substantially hurt science communication. Here, there may be some examples of the importance of strong graphics.

    Climate scientist Michael Mann’s graph of historical global temperatures might be an positive example where a technical graph helped the cause of science communication by breaking into public awareness and advanced political discussion.



    This visual made the point stick with people in a way that “unprecedented rapid change in earth’s temperature” never did. It was so memorable that even it got a nickname (the “hockey stick” graph).

    On the negative side, Edward Tufte’s short book, The Cognitive Style of PowerPoint (later incorporated as part of Beautiful Evidence), provided a compelling case that the default design of PowerPoint slides may have contributed to the explosion of the space shuttle Challenger.



    Tufte’s argument is that the design of the slides are so bad that they diverted attention from the deep problems facing the Challenger launch. While there were clearly a lot of management problems at NASA the contributed to the explosion of the Challenger and the death of her crew, the argument that this contributed isn’t easily brushed aside.

    External links

    PowerPoint does rocket science

    04 May 2019

    The Morrison billboard poster, part 2

    Since its debut on YouTube a few weeks ago, Mike Morrison’s suggested format has been getting a fair share of people trying it out. This one appeared in my feed via Hannah Hobson.


    I’m fascinated by this, because it shows how design can be like the whisper game. Things get changed with each repetition and you end up with something rather different (and sometimes incomprehensible) compared to the original.

    In Morrison’s original template, the entire content of the poster was contained on the left hand sidebar. The right hand sidebar is for extra supporting detail for superfans and ultraskeptics.

    This meant a presenter mainly has to help people on the left side of the poster. If the presenter is standing in front of it that right sidebar, it’s usually no big deal.

    In this version, the content is so much longer than the “structured abstract” that the content is split between the two sidebars.



     And to make matters worse, the break comes in the









    middle of the results section.

    As I have just demonstrated, unexpected breaks mess up your reading flow. Now you have to shift your physical location quite a bit to continue reading the results. You might have to dance back and forth between the start of the results on the left and the end of the results on the right to make sense of the whole thing.

    For the presenter, this version clusters audience members at both edges of the poster. The middle “take home” message becomes something of a dead zone. If there were meaningful content in the middle, you would walk people through in a continuous loop. But here, people have to jump over several feet.

    A good general guideline in graphic design is keep related content together.

    If you have so much stuff that you need two columns to explain it, keep those columns together.




    I’m also baffled by the QR code which says, “Take a picture to download the full poster.” Surely I can eliminate the middleman by just... taking a picture of the poster?

    Also, as noted before, the presence of a large take-home message in the body of the poster makes the title running across the top redundant. You only need one or the other, not both.

    Related posts