September 22, 2015
This is a visualization of the frequency of the words ‘hope’ and ‘crisis’ in the New York Times, between 1981 and 2009. The visualization reads like a clock, where 12pm is 1981 and 12 midnight is January 1, 2009. By Jer Thorp CC-BY.
Having good data visualizations in a paper or a research report can be useful not only to gain more attention of the general public, but also in making research easier to understand for your peers. Interesting visualisation, if licensed under Creative Commons can also be re-used by readers in their presentations, articles, or teaching materials, which may result in better dissemination of your research.
In the era of big screens connected to the global network we are getting more and more used to visually attractive content, which allows us to attain a lot of information at a glance. Especially in social media, multiple bits of data compete with each other and the most attractive ones get users’ attention for just a few seconds. Therefore, it is essential to insert an interesting and informative image to every blogpost or tweet. And because social media is at the heart of communication now, images have become really important. It can be seen as a generally negative process since valuable content can be missed in a noisy environment just because it is less colourful. On the other hand, our perception of the world was always been mostly based on colours and shapes, and good visualisation has bigger didactic capabilities than texts, even well written ones. Thus, researchers may benefit from the emphasis on visualization, demanded by a meme addicted society.
Data heavy texts, with a lot of numbers are especially difficult to read, even for specialists. Simple charts, or graphs can make data much easier to understand for everyone. However, not all data can be presented effectively in these forms. Sometimes, we want to expose several dimensions of data so that it can be spotted at just a glance. And here we are entering the realm of data visualizations. Having good data visualizations in a paper or research report might be useful not only to gain more attention of the general public, but also in making research easier to understand for your peers. Interesting visualisation, if licensed under Creative Commons or an equivalent license, can also be re-used by readers in their presentations, articles, or teaching materials, which may result in better dissemination of your research.
Art of data viz
One of my favourite examples of data visualizations are these made by Jer Thorp for the New York Times. They are 6 years old already, and they are much simpler than the later work of Throp, but they are both beautiful images and clear sources of information. They are visually interesting, elegant and easy to read. When looking at the circle graph presenting the frequency of the occurrence of the words “hope” and “crisis” in the New York Times (above), one realises what the stories are behind this data within a few seconds.
There are a lot of people specializing in data visualizations, and there is a huge demand for their work created by big commercial companies. But still a lot of researchers are lacking even basic visualisation literacy, which is a pity, since they are people who have the best stories to tell, and it would be good for all of us if they were able to tell them in a beautiful, elegant and effective way.
The people behind the Pasteur4OA project supporting UE member states in introducing Open Access policies, are probably the first in the open access advocacy movement to understand the benefits from data viz. This understanding resulted in a bunch of graphics visualising the The Registry of Open Access Repository Mandates and Policies (ROARMAP) data. These visualisations are not even half as stunning as Thorp’s ones (and they also include some mistakes), but they are probably a step in the right direction. Open access advocacy has to be data driven, clear and attractive, so it needs good data viz. If you have an idea for a visualization that is missing in Pasteur4OA, you can write about it on their publicly available pad.
But since there is a huge amount of open data, you can simply do it on your own. I am not an expert in visualisations, and I have little experience in coding (I am a sociologist, there were no programming courses in my studies programme) but I was able to make my first data visualisation within a day after I downloaded Processing. This is really simple programming language for visual design, intended to be easy to use even for those with no previous programming experience. Indeed, you can create a simple shape on your screen with just one line of code, and something as complex as map of the world displaying income inequalities with no more than 50 lines. For those who already know Java language, Processing will probably be super easy to learn since it is Java-based. Another popular tool for data viz is D3. I have not tried it but it is reported to also be easy to learn. So I think it is about time to think how to visualize the data from your next research project. And do not forget to publish your work under a license that allows re-use.