Practicing what we preach
How we applied iterative design principles in preparing our poster for Vizbi 2016
Francis Rowland and I gave a tutorial on Rapid user-centered design for data visualization in Vizbi 2016. In this tutorial we introduced hands-on techniques, common to user experience (UX) design, which help us to produce effective visualizations of scientific data. We emphasised the principles of trying to account for our audience’s perspective, of exploring different options, and of testing these options with actual users.
But perhaps it is interesting to explore how do these principles apply to more common challenges… such as preparing a poster for a conference?
Iteratively designing a conference poster
Many scientific posters are just long blocks of text with the odd graph in between (I still remember a friend from my postdoc years who turned his paper into a poster simply by changing the latex formatting command of his document). But, instead of just making a poster by pasting text from one’s most recent publication, a poster can be an opportunity to showcase one’s work in a more visually effective way.
We applied the iterative design approach that we described in our Vizbi 2016 tutorial in preparing a poster for the same conference. The poster outlines the User Experience (UX) work that we carried out for the web platform of the Centre for Therapeutic Target Validation (CTTV).
Identify the key message(s)
Our first step was to identify the key messages that we want to convey to our audience. We want to inform the attendees of Vizbi about the main problem that CTTV is trying to address and how the CTTV platform attempts to solve this problem. We also want to highlight how UX methods helped our colleagues who are developing the platform see this problem from the user’s perspective and work with each other and with potential users to design and test solutions in an iterative way. So our challenge was to present these key messages in a visually effective poster.
Who are the audience for the poster?
In parallel to identifying the key messages, we started thinking about the audience of this poster: Who attends Vizbi? What do they expect to see and what attracts their interest? Miguel Pignatteli, who is in our CTTV team, is a bioinformatician with a solid background in biology and strong visualisation skills. Miguel is quite representative of our Vizbi audience so we invited him to help us write the first version of the abstract for the poster which conveys our key messages. We also went through various images that we could use alongside text to communicate these key messages in the poster.
Review, and sketching alternatives
During this review, we also looked at previous poster submissions to Vizbi to get an idea of the conventions that previous authors had used. Equipped with an idea of who attends Vizbi and what they are used at seeing there, we sketched out 6 ways in which our key messages could be organised in a poster layout.
We reviewed these versions and selected three of them, from which we produced higher fidelity versions, to test with people who would represent our audience. The first version was a synthesis of two of the six initial layouts that we all thought would convey the key messages to our audience quite effectively. The second version was selected by Miguel who was representing our audience in this review. We chose as the third version the one that was most distinct from the other two. This is a common principle in considering alternatives to make sure that the various solutions span across a wider range of the design space.
The sketches with the six initial alternatives and the three options we selected to test are shown below:
Rapid idea generation, based on defined goals
Up to this point, we had spent less than 1.5 hours from the initial formulation of the key messages, to our review with Miguel which gave us the first version of the abstract (our main textual content), images to pair with the text, and three out of six initial layouts to test out with our audience. Our experience of working collaboratively like that in this and other occasions is that it can be quicker than working without clearly defined goals and in isolation from our audience. It also helps us come up with outputs of higher quality than what each of us would produce by working on its own.
Making 3 prototype posters
I took three steps to make the 3 versions to test with our audience: First, I formulated the text for the poster, which was largely based on the abstract. Then, I placed the text on each poster, making a outline without any images. Finally, I added the representative images:
This took the best part of an afternoon and we ended up with three draft posters with the same text and images but different layouts:
Testing the prototypes
Luckily, the EBI is full of bioinformaticians interested in data visualisation who come together every day at 3pm for our departmental coffee. To test the three posters, we displayed them in the area where the coffee is served and gave our colleagues a chip to vote for their favourite poster. Within 30 minutes, we had a clear winner, which interestingly wasn’t my own favourite.
In addition to raising awareness about the UX work we had done for CTTV, this exercise also demonstrated to our colleagues one of the main general principles of UX, namely to try to come up with various alternatives and to test them with users before committing to a particular solution. While our colleagues were voting, we heard some discussion about how we had managed to come up with not one, but three versions, very quickly.
From our perspective, this was a useful, simplified example of rapid, user-centred design.
Acting on feedback
This process also gave us the opportunity to collect additional feedback from our colleagues. One of the most striking comments was that there were no women in the photographs of the posters. This was unintentional given that there were many female participants in the research, design and testing activities that we conducted for CTTV.
For the final version of the poster we addressed this apparent gender unbalance and also tweaked the text based on the feedback from our colleagues.
Overall, our poster for Vizbi took less than 3 days from conception to final submission and was a great opportunity for us to practice what we preach for the common challenge of designing a poster for a scientific audience!