Kamis, 16 Juni 2016

Latest modern science | Search engines for technical graphics - Si Bejo Science

Add Comment
How important are academic graphics? A new pre-print in arXiv argues, �Pretty damn important.� This news summary of the technical article says:

(T)heir most remarkable discovery is that the most successful papers tend to have more figures. By plotting the number of diagrams in a paper against its impact, the team concludes that high impact ideas tend to be conveyed visually.

Lee and co say there are two possible explanations for this: �That visual information improves the clarity of the paper, leading to more citations, and higher impact, or that high impact papers naturally tend to include new, complex ideas that require visual explanation.�

The team has a search engine for scientific graphics called Viziometrics. My first pass, for �crayfish,� gave a mess on non-intuitive results (click to enlarge):


Things improved markedly when I selected only for diagrams and photos, however.


Speaking of searchable graphics databases, Atlas looks promising for some purposes. I tried searching for something that I thought must get plotted a lot in science, �Impact Factor�:


Nothing looked relevant to scientific publication, so I tried a couple of other topics familiar to me. I had success with �lobster�, because I reckoned there would be fisheries data. There was:


Things get good when you drill down to a single graph:


There�s a reference, so you know where the data came from. You can download the image created by Atlas, or download the data yourself in a plain text (CSV) file. Atlas is a product of the Quartz online news outlet. I�m not sure yet if it only includes data from Quartz stories.

These are not going to replace Google Images or Flickr any time soon, but they might be useful for some purposes.

Hat tip to Ananyo Bhattacharya for Viziometrics and to Knight Science Journalism for Atlas.

Kamis, 09 Juni 2016

Latest modern science | What is the �ePoster� format? - Si Bejo Science

Add Comment

I�ve been predicting that we�re going to see a slow decline in paper posters for a while, so I was interested when John Coupland and Lady Scientist drew my attention to the joint annual meeting of (deep breath)...

  • The American Society of Animal Science (ASAS)
  • The American Dairy Science Association� (ADSA�)
  • The Western Section of the American Society of Animal Science (WSASAS)
  • The Canadian Society of Animal Science (CSAS)

They are having presentations in an �ePoster� format. Their instructions are here. I can�t quite visualize this yet, but as near as I can tell, it is an illegitimate love child of a PowerPoint slide show and a paper poster.

It�s the size of a poster (about 40 inches wide)... but you can have multiple screens of information, with hyperlinks and videos (like slides). The conference FAQ says:

On average, presenters will normally have 3-�5 pages of content on their e-�posters.

How is this different from a slide show? The conference organizers are squelching the �Click to advance� button, and replacing them with timed animations and navigation buttons.

I am skeptical of this hybrid format. The single sheet of a paper poster enforces discipline. You have to make hard decisions about what to include or not include. I worry that allowing multiple slides will mean that people will upload their standard PowerPoint deck and just give their standard talk repeatedly instead of once. As far as I can tell, the �normally 3-5� comment notwithstanding, the only limit to the number of slides is the total size of the PowerPoint file.

On the other hand, presenters do have to upload the poster in advance, so you won�t see PowerPoint decks that were cobbled together on the plane trip on the way to the conference.

The conference provides two templates. Here�s a sample of one (with instructions intact):


And another, which I think is less effective:


I�m not a fan of these templates. For one, the blue background and text are fairly low-contrast. Because these are on a monitor, white text will be more effective than black: it will glow from the light behind it. The columns on the first seem rather wide. Displaying all the sections on the second seems rife for confusion.

And a conference logo. In prime real estate. I know what meeting I�m at, I�ve already paid the fee, there�s no need to advertise it any more.

Based on the conference documents, this is all being run by a company called ePosterboards, who are new to me. One of their services is they will provide design assistance to poster authors, for an unspecified fee. You have to contact them for a quote. This makes me super curious as to how much they are asking for.

Kamis, 02 Juni 2016

Latest modern science | Critique: Notorious DRG - Si Bejo Science

Add Comment
This week�s contribution is from Zach Sperry, who gave me permission to share his poster from the 2015 Neuroscience meeting. Click to enlarge!


Nobody should be embarrassed by a poster like this. The core design of this poster is solid. It�s a clean, three column layout that leaves no doubt as to how it should be read.

But... there is a lot going on in this poster. It might have benefited from the four tips on shortening posters I had just a few weeks ago.

Things I might do:

Take the university and lab logos in the title bar out. This would allow you to shorten the author and institutions credits from five lines to maybe two, and make the title bigger.

I cannot emphasize this enough: at big meetings, your title needs to be visible from the moon. Big meetings set poster boards far apart to have aisles for people to walk in. And Neuroscience in the biggest of the big. Do not skimp on space for your title!

As journalists say, this poster has buried the lede. The �Goal� statement is crystal clear, but it�s buried at the bottom of the first section, and the italics are not enough of a signal to show its importance. I like the clarity of the �Goal� statement so much that I might just hack that whole section down to that one sentence.

Try shrinking everything by 5-10% and increasing the white space between each individual element.

The text blocks are quite dense and dark. The typeface appears to be plain ol� Arial. I might try a thinner typeface and less bolding to make the text blocks look lighter.

There are quite a lot of bright colours on the poster, with red, green, orange, and blue all making appearances. While the area they cover is small, which generally favours those more intense colours, there are still a lot of them, which contributes to the feeling of business. They make sense in the graphs on the right, but the intense red for labels in the central bottom figure, or to make outlines in the right figure (�Cell bodies (blue pixels) transformed...�) might be a little too much.

Like the introduction, I would like the conclusion to be much tighter. The first of those three paragraphs alone might be enough. It tells you the two key take-home messages:

  1. The recording worked.
  2. We got new information from this recording.

Kamis, 26 Mei 2016

Latest modern science | Link roundup for May 2016 - Si Bejo Science

Add Comment
Matthew wins for best poster design cartoon this month:


You can see more of his cartoons at Errant Science. His inspiration for this one?

Just looked at a draft of a poster, there was text in size 2 point�

Just to drive the point home, let me say: Writing text in a 2 point font on a poster is dumb. Do not do that.

James Hamblin, writing for The Atlantic, has a fascinating article about using colour to increase the readbility of text. Here�s an example, where colour is used as a cue to tell you what the next line is:


This specific example is from a company called BeeLine. They have plugins for Chrome for the web and PDFs. Here�s this blog viewed in Beeline:



I haven�t had success with reading my PDF reprints in BeeLine colour yet.

The Atlantic article suggests there are many more possibilities to improve the reading experience beyond what we have learned from the printed page. I would not recommend trying it for a poster quite yet, because the unfamiliarity might be confusing or annoying for readers. A browser plugin is not like a poster: the reader has the control in the former, but not the latter.

The Atlantic article also mentions the Microsoft typeface Sitka (sample at right):

(Microsoft researcher Kevin Larson�s) team also recently launched a new font that was designed for the best possible readability. Called Sitka, it went through a multistep, iterative design-test process. Each letter was changed and adjusted to maximize ease of reading � as opposed to most other fonts, which are made to mimic typefaces that existed in print media. �Times New Roman was designed to work very well with the technology of the era,� Laston explained. (I asked him if he has, then, created the most legible font in history. He said he �wouldn�t go that far.�)


I might just try Sitka on my next poster.

I really enjoyed this blog post by Stephen Few about 3-D graphs. You know, like this one:


Not only is the article thoughtful, there are some great one-liners:

(P)oking holes in Edward Tufte�s work in particular now qualifies as a competitive sport.

And I like the conclusion:

It is important to realize that what is often claimed by infovis researchers is just plain wrong, due to bad science. I wholeheartedly agree... that we should not accept any data visualization principles or practices as gospel without confirming them empirically. However, we should not throw them out in the meantime if they make sense and work, and we certainly shouldn�t reject them based on flawed research.

Today in �Colour is a subtle thing,� Ed Hawkins looks at how �rainbow� coding for maps led to some incorrect interpretations.



Hat tip to Rob Simmon.

I�ve followed the fate of �dynamic posters� at the Neuroscience meeting since they were first announced. I think a fee to have one is new:

$150 fee to present a Dynamic Poster #SFN16? pass. - Drugmonkey

And the question remains how many people are genuinely using the format to its fullest:

80% of those are just people who wanted to give a talk. Rarely actually need video. - Dr. Becca

Michael Hoffman is looking for new graphics software:

Wish I had something that made it as easy to make diagrams as PowerPoint but still publication quality like Illustrator.

In the replies to this tweet, people bring up xfig, Sketch, Drawio, Omnigraffle, Canvas, Graphic, and Pages.

QUote of the moment:

�Typography is frozen sound.��Ran Zheng

Hat tip to Ellen Lupton.

Another university, another new logo freak-out.


Kamis, 19 Mei 2016

Latest modern science | The problem with point size - Si Bejo Science

Add Comment
�What�s the smallest point size you can put on a poster?�

This is a common question, but it�s not one that has a simple answer.

I know many scientists read this blog, and scientists work in a world where measurements are universal. 37�C is 37�C no matter wherever you are, whatever you�re measuring, and are exactly comparable. Someone from a scientific background probably thinks that two identical pieces of text � in different fonts but the same point size � should take up the same space.

I have bad news. Point size does not work like that.



These two paragraphs are both set in serif typefaces, both nominally 16 points in size, but one takes 19% more column length than the other. This difference can arise because individual letters in the two fonts might have the same height, but different widths. The letter O may be a wider circle, or a narrower oval, for example.

That 19% will make a big difference in your layout, even if the two blocks of text are similarly readable.

I have selected two fonts with a fairly large difference here. Many other standard fonts will probably be more similar in their use of space. But it points out that you can�t rely on font size alone to guide your poster design.

Instead of blindly following a minimum font size, work from a couple of guiding principles.

  1. The bigger the text, the better.
  2. Test, test, and test some more. Print full sized sample paragraphs at the point size you want to use (12 point 18 point, 24 point, 30 point), tack them to a wall, and stand back a couple of meters and see how they look.

But if all thoughtful design and testing stuff bores you, the answer is:

Nothing smaller than 24 point on your poster.

There. Happy?

Kamis, 12 Mei 2016

Latest modern science | Four simple tips for shortening your poster - Si Bejo Science

Add Comment

Few things will turn away a potential poster viewer like long paragraphs of text. So one of the recommendations I (and many others) make for posters is to write less stuff. But it is not easy.

There�s a saying (wrongly attributed to Abraham Lincoln), that if you have a short time to cut down a tree, spend most of it sharpening the axe. Here are some ways to sharpen your editorial axe.

1. Walk away.

When you�re in the middle of a project that you designed and carried out, everything seems important. But time away from something helps bring clarity. Think about a favourite album or TV series that you haven�t watched in years. You won�t remember all of it; you will remember the highlights.

You can get clarity by not working on a poster for a few days, then coming back at it with fresh eyes.

I think this is be the surest and best approach, the problem is that it takes time. You have to start early, and allot �cool down time� of a few days where you do not look at the poster. Given how many academics don�t want to give posters because they want to slap together a PowerPoint talk on the plane on their way to a conference, getting them to work on posters well in advance is a tall order.

2. Show it to someone else.

An outside viewer doesn�t have that emotional or intellectual investment in a project that you have. The further away you can get, the better. Show your poster to someone who isn�t in your lab. SHow it to a non-expert. Show it to someone with a different skillset.

Just remember that an outside observer is not necessarily an unbiased one. Everyone has their own tastes and preferences and styles. An outside observer may not be objective, but they will at least have different biases than you.

3. ABT.

�ABT� is short for �And... But... Therefore.� You take two facts (joined by �and�), followed by the complication (�but�), and a resolution (�therefore�).
 
It is one of the single most effective tools I have found for drilling down to a key point. And it has the advantage of being quick (unlike #1) and not needing others (unlike #2).

I�ve done this with many poster presenters. When I ask them to talk about what there poster is about, it often takes a few minutes. I don�t think many of them believe me when I say they should be able to summarize their poster in a sentence. Then I do it using the ABT format. And I can usually see the expression on their faces indicating I�ve hit very close to the mark.

I wouldn�t recommend condensing the entire poster to one sentence, but it�s great at chopping a couple of lengthy introductory paragraphs into one crisp sentence.

Randy Olson first introduced this sentence structure in Connection: Hollywood Storytelling Meets Critical Thinking (which I reviewed here), and has continued working with this tool in Houston, We Have a Narrative.

Additional: Randy notes that you can learn more about ABT in Story Circle Training here. He also advises for the verbal presentation that goes along with the poster:

1) Say your ABT, 2) Ask what person studies, 3) Find bridge between the two (from Samantha Roy)

4. Practice ruthlessness in all your writing. 

The Elements of Style by Strunk and White is not a perfect book on writing. Likewise, the fretting about Marxism in George Orwell�s essay, �Politics and the English Language,� is very out-of-date. But both remain worth reading because of their emphasis on being concise.

There are many lists that alert you to lengthy stock phrases that can be replaced with shorter words. Once you attune yourself to stock phrases (�At this point in time,� �The fact that�), it becomes easier to recognize them, cut them out, and replace them without losing any meaning (�Now,� �That�).

External links

To Cut Down a Tree in Five Minutes Spend Three Minutes Sharpening Your Axe
Connection: Hollywood Storytelling Meets Critical Thinking review

Kamis, 05 Mei 2016

Latest modern science | The inverse colour intensity rule - Si Bejo Science

Add Comment
Colour on a poster is powerful, but can be difficult to use well. One of the problems I often see is that people create posters with deep, bright, saturated colours that cover large areas:


It�s hard to get the full effect of this on a small computer screen, but those bright colours on a poster that several feet wide can be like a punch in the eyes. Size signals importance. So do bright colours. The combination of both can be overwhelming and is rarely necessary.

The larger the surface area you colour, the less you have to do for it to �read� as a distinct colour. You can go to a very light, almost pastel shade, and it will still come across as clearly distinct from a background or other colours. People can readily tell that you�ve highlighted something in blue.


If you have something that is very small like a select point of data on a graph, or something for emphasis � that�s the place where you can use those bright, intense colours to draw attention. Subtlety won�t cut it if something is small:


Bright colours can draw attention to something that isn�t dominating by virtue of space.


And the moral of the story is: The bigger the space you colour, the less intense the colour should be. I�m still not sure what the name of this principle should be, but �Inverse colour intensity� will do for now.

(The text? Happy Revenge of the Fifth, everyone!)
Langganan: Postingan (Atom)