A central component of the Digital Humanites internship project requires that I develop a research paper on the current trends and tools in the field of data visualisation. My initial reading suggests that current research and development in the field is dominated by the voices of Alberto Cairo, César A. Hidalgo , Ali Almossawi and John Grimwade. Their writing and data visualisation products make for intriguing reading and exploration. I am drawn to the beauty and magic of data visualisation and find myself spending hours simply experimenting with the possibilities inherent in projects like Hidalgo’s Pantheon, a visualisation tool for the study of collective memory, languages and technology.
The writing on data visualisation is equally compelling, Almossawi and Hidalgo’s magazine article in The Scientific American draws an interesting analogy between the 17th century astronomy discoveries of Galileo to today’s data visualisations. The authors describe data visualisation as a virtual telescope, enabling the user to find a way through the deluge of big data. They suggest that “These tools allow us to explore the fluid landscape of bits, instead of the rigidity of atoms, giving rise to a new medium that is helping us comprehend the complex while simultaneously providing a new means of artistic expression.” The image is nicely completed with a depiction of users changing from the role of spectators to explorers. Hidalgo is a fascinating talent in the realm of data visualisation, amongst other fields, as the video below explores.
He is involved in the creation of Data Visualisation Engines, online tools that allow people to work with the visual aspects of a data set. He is the creator of the previously mentioned Pantheon, and of DataViva, a tool involving one billion visualisations for the study of trade, employment and education in Brazil. These are exciting and transformative tools, Hidalgo’s work flow in the creative process is itself compelling viewing, in some cases involving preparatory sketches on whiteboards, see figure 1 below. As an aside, the use of paper and pencil, or in this case whiteboard and marker seem to be integral to the end digital product for many data visualisation developers, evident again in figure 3 later in this post.
Figure 1 César A. Hidalgo Preparatory sketches, MIT Media Lab video.
Arthur C Clarke’s maxim that any sufficiently advanced technology is indistinguishable from magic comes to mind when, in the course of my research, I become transfixed by the beauty, magic and wow factor of some of these data visualisation tools and realise I have wandered far off task! The project timeline forces me to adhere more strictly to the requirements of the project at hand, a data visualisation of the poetry of John Donne in 17th Century manuscripts. I narrow my focus to the discovery of a tool that can efficiently and simply present and explain the data on Donne’s poetry and on the ways in which it can be significant. The more sophisticated data visualisation engines require a certain level of mastery of the underpinning technology. For the project at hand, I maintain that a simpler visualisation tool will suffice. It is important that it is a tool which will not require the user to invest time in the mastery of a complex technology before they can attempt to decipher the data.
John Grimwade warns against this, apparently common, tendency to fall in love with the magic and complexity of the visualisation tool, rather he stresses the importance of placing the information in context. He suggests that this contextualisation is done with annotations rather than making the assumption that the user will be happy to take on the role of editor and spend time teasing out a complex data visualisation in order to figure out the meaning of the data.
This clarity of meaning as a defining feature of a successful data visualisation is a theme expanded upon in Albert Cairo’s profile of Grimwade in The Functional Art. Here Grimwade and Cairo discuss the the failure of data visualisation developers to prioritise clear communication of information rather than the development of a beautiful but complex tool requiring the mastery of technology. Instead, they see the purpose of visualising data as primarily to explain or describe a topic by revealing patterns, from which the user can easily draw conclusions. Again, it is to facilitate further explorations from these conclusions rather than to turn the user into a data analyst.
This point is usefully demonstrated by investigating Grimwade’s creative process in developing a data visualisation or infographic as outlined in The Functional Art, see figure 2.
Figure 2 Grimwade’s Transatlantic Superhighway reproduced in The Functional Art
In the flow of developing the visualisation above Grimwade worked as a reporter but also as an editor, annotating and placing the information in context. His work in process is not born with digital coding but rather involves using pencil and paper for sketching and note taking in an effort to isolate the key elements of the topic, as demonstrated in figure 3.
Figure 3 Preparatory sketches for the data visualisation depicted in The Functional Art
Grimwade, in conversation with Cairo discusses the creation of “amazing interactive tools with tons of bubbles, lines, bars, filters, and scrubber bars” but steers the reader to
“Think of Hans Rosling and the way he interacts with his wonderful bubble visualizations. He doesn’t just show stuff; he explains the main points, focusing the reader’s attention on the most interesting parts of the information. After that, if readers want to navigate deeper into other possible stories, they can do it. But first, they are exposed to a traditional, linear narrative that lays out the basic facts.”
In my my last blog post , I discussed the problems inherent in not being sceptical of data visualisations, in accepting the information they contain more blindly than we we would do if the same information was presented in text format. At this point in my project work, I am concerned with not being seduced by the magical possiblilities inherent in data visualisation tools and their “tons of bubbles, lines, bars, filters, and scrubber bars”; it is time to once more focus on the project timeline.