From Data Sketching to Dataviz: An Experiment with The Vistorian

In this post I do not mean to give a complete overview of the Dataviz workshop in Aix last week (February 14-15). Instead I will give a personal account of the hands-on session we had with Benjamin Bach (Edinburgh University) on Friday.

During the hands-on workshop, participants had the opportunity to bring their own data, to play with them and to experiment with uncommon visualization tools. I personally chose the Shanghai Rotary Club as a case study and I used the data I recently compiled from three successive rosters (1930, 1940, 1948) that I found in the Calder Papers at the Hoover Institution Archive (Stanford).

But before diving into digital techniques, Benjamin invited us to leave our computers aside for a while and to just use a basic pen in order to sketch out our ideas on a simple sheet of paper. This is what he calls “data sketching”.

Data sketching

As Benjamin rightly put it, data sketching is critical in the ideation process. It also helps to explore the visualization design space; to familiarize ourselves with our data; to structure our data and messages; to make informed choices about your visualization design; and to communicate ideas.

In our data sketching session, we followed three steps:
  1. We first divided a paper sheet into six sections (see picture 1 below) : in the first section at the top, we had to write down data elements, attributes and relations ; on the third and  fourth sections below, we mentioned the research context and our intended audience (conference, research paper, website, teaching…); at the bottom, we articulated our research questions and the messages we wanted to communicate;
  2. Using new sheets, we had to sketch various possible ways of visualizing our data with a view to answer our research questions (see picture 2)
  3. On the last sheet, we had to refine these proposals in order to elaborate our final visualization after exchanging with other participants (see picture 3).

At the end of the data sketching session, I came to the conclusion that while basic tools and common visualizations (Excel tables and graphs) may be sufficient for preliminary statistical analysis (members nationality, positions, class of business), it was not satisfying for analyzing relational, spatial and temporal patterns that are not immediately legible in my data.

Exploring tools

After the data sketching phase, we had the opportunity to experiment with various tools that Benjamin introduced. Personally I was not really convinced by Datawrapper or Raw Graphs (not adapted to handle complex data) and I resigned myself not to using D3 (although it is really flexible and powerful, it requires programming skills and implies a long learning curve) and SNA tools that often proved deficient in handling space and time.
I finally turned to The Vistorian – a tool that Benjamin Bach has been developing for about two years – which I particularly like because of its user-friendly interface and because it is convenient and quick to handle. More importantly, it allows to manipulate massive and complex data (my table had only 325 rows but one of my colleagues had several thousands of them and it worked nonetheless, as shown below). Last but not least, it is specifically designed to visualize space and time. I found it easy to upload my data as a simple csv file, and easy also to select the dimensions I wanted to visualize (using the unfolding bar) and the ways I wanted to display them (color/link type, node size, edge width). Once the data are uploaded, one can choose between different views (node link, adjacency y matrix, time arcs, maps) and it is very easy to switch from one to the other so as to explore different ways of displaying the data.

Visualizing historical big data with The Vistorian

As a historian particularly concerned with chronology, what I prefer is the time slider that allows one to navigate across time and to see patterns of change in the data – in my case, I can explore how Rotary membership evolved over time – in terms of importance, nationality, occupations or business classification (I know but a few tools that offer this functionality).

The “node link” view in The Vistorian

I found, however, that The Vistorian may be slow at times and we sometimes need to close and restart the application to be able to use it again. Moreover, it is not always easy to navigate between the original table view and the various views. And it would be wonderful to have more customizing options (e.g. to change the color of nodes or play on their length). But this is a very partial experience of The Vistorian. I have presented but a very limited sample of the available options and my preliminary explorations still deserve further examination. This short-sighted overview of this exciting tool certainly does not do justice to its manifold possibilities (notably the adjacent matrix and mapping views) and I encourage other historians to try it and “play” with their data.

Wrapping up

What can we learn from this dataviz workshop?
  • First, data sketching may be a useful exercice even – or particularly? – in view of “serious” scholarly purposes (e.g. publishing articles in academic journals, presenting a research papers at a conference).
  • the workshop has offered a rare opportunity to exchange with colleagues, to communicate our ideas (which is not always easy), to receive critical feedback on our own research, in order to better understand what doesn’t work and why it doesn’t work  and to imagine alternative ways of using our data (e.g. my teammate suggested that I should use treemaps to visualize social-professional categories and sectors of activity instead of the usual bar or pie charts) ;
  • It is good sometimes to return to the foundations of visual thinking before venturing into sophisticated visualization techniques. While a lot of self-claimed “revolutionary” tools today are good at marketing themselves, they are pretty much limited for conveying scientific messages. Through data sketching we become more critical toward “off-the-shelf” applications – and it appears that too many of them tend to forget the basics of visual communication and graphics semiology, as they have been remarkably established by Jacque Bertin as early as 1967.

Keeping this in mind, data sketching and experimenting with tools like The Vistorian invite us to go beyond tables and common graphs (bar and pie-charts) – and more generally beyond numerical or database-driven way of thinking – to move toward more creative visualizations.

It is not so easy for classically trained historians to think visually. Yet the effort is worth trying. Sketching, drawing and visualizing historical data are not a waste a time. Visualizations allow to see things we would not see through a linear reading of sources, to reveal hidden patterns and to imagine new ways of analyzing and interpreting our data. It may help eliminate wrong hypotheses and raise unexpected research questions. In a word, it is an opportunity to thing differently.


Cécile Armand

Ancienne élève de l'Ecole Normale Supérieure (ENS) de Lyon (2006), agrégée d'histoire (2009), docteure en histoire (2017), postdoctorante à Stanford University (2017-8) puis Aix-Marseille Université dans le cadre du projet ERC “Elites, Networks and Power in modern China” (2018-). Ma thèse soutenue à l’ENS Lyon en juin 2017 proposait une histoire spatiale de la publicité à Shanghai (1905-1949). Mon nouveau projet porte sur l’invention du consommateur et de l’expertise sur le marché en Chine républicaine. Au-delà, mes centres d'intérêt couvrent l'histoire urbaine, spatiale et sociale, les outils et méthodes numériques en sciences sociales, l'historiographie et l'épistémologie en général.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.