Kamis, 28 Agustus 2014

Latest modern science | Link roundup for August 2014 - Si Bejo Science

Latest modern science | Link roundup for August 2014 - Si Bejo Science

�Only one idea.� Although this post from Terry McGlynn is about giving oral presentations, I think there is something to apply to posters, too:

When I gave the talks that (at least I suspect) kicked butt... They were all short on data and they only had one idea. ... I intentionally pared back on the data and the overarching research agenda. I just wanted to speak to an idea and wasn�t too concerned about how it made it me look. And, it turned out, the result was that people thought I gave a really good talk.

Kamis, 21 Agustus 2014

Latest modern science | Critique: Megafauna - Si Bejo Science

This week�s poster is from Benjamin Seliger, and is used with his permission. Click to enlarge... or perhaps I should say, �megasize.�


I would stop at this poster if I saw it at a conference. There is much to like about the design. The visuals are very strong and very prominent. I love the pictures of the megafauna, the plants, and the maps. There is not too much text.

This poster accidentally demonstrates the power of proximity and white space. When I glanced at this poster, I thought, �This is a very nice two column layout.� But I should have thought, �This is a nice two row layout.� This poster is meant to be read across first, not down, which is the opposite of what I thought from a glance.

I am supposed to see the poster elements in this grouping:


But instead I see this grouping:


The problem arises because there is a wide, generous margin between the columns, but almost none between the rows. We group things that are close together. The authors have tried to signal that these are in rows using horizontal dividers, but the �signal� from the wide margin in the middle is completely overpowering that from those skinny little lines.

That the headings are not that much bigger in size than the subheadings is not helping matters. Compare the size of �The data� to �Joshua tree� underneath it. The �Joshua tree� and �Honey mesquite� subheadings are reinforcing that this is a two column layout instead of a two row layout. I also wonder if flipping the text position (above the animal picture, but below the plant picture in the top row) is contributing.

Fortunately, the solution is simple. Make the margins between the rows bigger than the divisions between the columns. Here�s a quick and dirty revision:


Margins beat lines and boxes in signalling the organization of your material.

Kamis, 14 Agustus 2014

Latest modern science | Critique: Protein simulations - Si Bejo Science

I was asked to look over this poster (click to enlarge). Now, this is a draft poster, so some of the large empty spots are deliberately empty, because the data were not in when the draft was tweeted to me.


Have I mentioned lately how much I hate photographic backgrounds on posters? I can�t recall ever seeing one that was effective. If the image is good enough and recognizable enough to show, why would you cover is up with text and data?

The gray photo background picked here is a recipe for disaster. The text is hard to read already, and the gray background will make it almost impossible to read at an distance. It�ll be worse if lighting is dim.

Even if the bothersome background is removed, I doubt this poster will pass the arm�s length test. The main text � and to a lesser degree, the headings � on this poster would both benefit from being bigger. This might require some judicious killing of darlings, but the poster will be better for it. One thing that might help is to make the acknowledgements smaller (make it �fine print�) so the conclusions can be bigger. The conclusions are more important, and one way to signal that is how much space it takes up on the page.

I don�t know what�s going on with that blue thing in the middle of the title bar, apart from it distracting me from the title, and making the title harder to read.

The colours in the methods flowchart seem to be picked almost at random. Kuler is a very useful too for picking harmonious colours.

While I can�t tell for certain with this low res image, but it looks like a lot of things aren�t aligned. The heading boxes definitely do not line up with the images in the lower left.

Kamis, 07 Agustus 2014

Latest modern science | Critique: P7C3 - Si Bejo Science

Bhavna Guduguntla asked if I could help a friend with this poster (click to enlarge):


A few weeks ago, I mentioned that your title (headline) is 90% of your communication effort. This poster would benefit greatly by taking that one board. The title here is probably one of the least visible things on this poster, for two reasons.

First, it�s competing with a bunch of logos, which are sitting in the poster�s prime real estate: the upper right corner.

Second, the title is thin white text on a light gray background. Those facts alone make the title inconspicuous, but it�s made even worse because it�s surrounded by several black elements: one of the logos, and the authors� last names. Your eyes are drawn to the highest contrast elements, and it�s not the title. And darn it, you should see the title.

I like the way the authors� last names are set in black. The last names on a scientific poster are important, because scientific papers are normally referenced by last names. If the names were properly subordinate to the title, this would be an nice design choice.

The main text looks crowded and ill-chosen. The boxes have so much text that the words seem ready to burst out of their boxes. Then, the text is sometimes centered, and sometimes left aligned. Consistency always helps give the appearance of considered, ordered decisions, which is what you want for a research poster. Stick to one format for all text!

The molecule between the two columns is distracting. It sits uncomfortably between the introduction and the data, but doesn�t clearly belong to either section. It�s also crushing up against those other sections.

There are two columns in the middle of the poster: �Log� and �Activity�. I am wondering if these are supposed to be one table? If so, they should touch, and not have a solid band of the background colour between them. Tables are generally not the best way to show data, and I wonder if there is any way to show that as a graph.

I am worried that the graphs at the bottom ones will be hard to see. They are on a coloured background, which reduces their contrast and visibility immediately. And the situation is made worse by the very fine lines used for the graph.