Conference Presentation:

Finding Lost Stories in the Digital Morass and Emerging Modalities for Telling Them

This page supports a presentation that I gave in Denver, Colorado on July 9, 2011 at the Third Applied Legal Storytelling Conference. The presentation presented a work-in-progress providing a brief history of the development of visual storytelling, and adressed how computer-generated visualizations of complex data are developing to manage the massive digital streams we currently generate.

In an effort to save cost and be more eco-friendly, I prefer to distribute most of my supporting materials through a web page, rather than in paper format. In addition, this form of distribution allows me to link to examples and other sites that I showed during the presentation, in case attendees would like to learn more. I hope these resources will be helpful to people who attended the presentation, but also to those who were not able to attend.

Prezi:

Here is a link to the Prezi that I went through during my presentation. I have been using Prezi more and more for these sorts of presentations because it breaks the linearity of PowerPoint, which I find limiting. Using Prezi also allows me to provide what Ed Tufte calls a "SuperGraphic" - a picture of the entire presentation - as a handout. He recommends making a SuperGraphic, and handing copies of it out before your talk, and I tried that for the first time with this presentation.

Links to Graphics discussed in the Presentation: 

In the presentation there were numerous examples of graphics, starting with a very simple one illustrating the concern about "information overload." Next, I noted how we have a long tradition of visual storytelling, from the Lescaux cave paintings in France, to Egyptian hyroglyphics. I noted that we have visual storytelling in maps, from Durer's engraving Celestial Map of the Northern Sky, made in 1515, to the renowned NY Subway map by Massimo Vignelli. I also noted that tattoos and logos are examples of visual storytelling, such as the Amazon logo (smiling face from a to z) and the FedEx logo (with the arrow).

The next step of the progression is from visual storytelling to visual representations of data. The classic of the genre is by Charles Joseph Minard, who created a graphic in 1869 that tells the story of Napoleon's march to Moscow in 1812. This graphic contains a great deal of data on troop levels over time, as well as temperature at various stages of the retreat from Moscow. Ed Tufte calls this graphic "probably one of the most perfect ever made" and says that it was "the first anti-war poster ever made." A similar poster provides a "cyclogram" (a diagram showing cycles of a natural orbit) made by Russian cosmonaut Georgi Grechko during the 96 day voyage of Salyut 6. It shows, in meticulous detail, 1500 sunrises and 1500 sunsets, as well as much other data about the mission. These two posters are great examples of representing a great deal of data in a quickly understandable form. But, they contain a manageable amount of data.

As data gets more complex, variant, and in different formats, the question arises as to how we might create similar visualizations of complex data. One example is the Radiation dose chart, which offers up a great deal of data gleaned from many sources (cited in the graphic) in an easy to understand format. Another is the Billion Dollar O Gram, created by David McCandless. As many of us are, David was frustrated by the many "billion dollar" figures that we hear about in the press, whether representing war costs or the valuation of certain companies. He thought it would be helpful to create a graphic that offers comparisons between these giant figures. And, indeed it does help. These two graphics - the Radiation and the Billion Dollar - show us emerging modalities (or semantics) for connecting signs and chart formats with complex data.

An intriguing question - and one that this presentation lead to - is what happens when the data is so big and so complex that these sorts of visualizations are - effectively - impossible to create? We run the risk of losing information - some of which might have been assembled into a story of great import in a legal case. The answer to this question is being discovered and developed in computer science departments around the world (the "emerging modalities" of the title of this presentation). In simple terms, new techniques are being created to generate visualizations from massive data stores. These are visualizations designed by the user but generated by the computer from vast amounts of data that would otherwise (effectively) be unusable. A simple example of such a CG Visualization is a Wordle. Wordles are visual diagrams showing the frequency with which certain words are used - the larger the word, the more it was used in a body of text. Here are two Wordles created from toy advertisements aimed at Boys, and another aimed at Girls. Very quickly, you can learn something important from a vast amount of data (the many words used in many advertisements that were the subject of this study) by simply looking at the Wordle for a few seconds. Here is another CG Visualization of Twitter traffic during the World Cup in 2010, and here is another, which analyzes the use of swear words in Twitter feeds across the country.

Perhaps the most interesting, and most useful to attorneys, is the use of CG Visualizations on large stores of E-mail over time. Over a million E-mails sent within Enron prior to its meltdown were published by the SEC on the internet, and since then, computer scientists have been studying ways to effectively create visualizations of connections and relationships between the various players in that case. Here is a visualization of the Enron E-mails. Other scientists are developing ways to rank the relative "importance" of certain emails over others, producing visualizations that look like this.

The next step is to make these sorts of CG Visulizations accessible outside the lab, and Hans Rosling is making great strides in this. During the presentation, I showed this video of him explaning the Gapminder data project he and his students have been working on.

And so where does that leave us? I think it brings us to a hopeful place, where even though data streams are increasing, tools are being developed to use them and learn things from them that would otherwise be lost. This is good, because increasingly, lawyers will have client stories to tell that are currently being lost in the data stream. With these tools, they should be able to find them, and tell them.

Links to Examples referenced in the Presentation:

For more information on these themes, look to: Ed Tufte, David McCandless, Hans Rosling, Good, Computational Legal Studies, and, just for fun, go create a Wordle.

The Third Biennial Applied Legal Storytelling Conference took place at the University of Denver's Sturm College of Law, on July 8-10, 2011.