[infosthetics@strataconf 2011 by guest blogger Collin Sullivan]
Day 3 at the 2011 Strata Conference seemed much less thematic than Day 2, but was every bit as interesting. It began with some of the most engaging keynotes of the week, focusing largely on just what is possible with data ("Making Data Work," as it were). Simon Rogers of The Guardian newspaper (Free Our Data: The Guardian's Approach, video, pictured above) emphasized the utility of making data available to the public. He explained how he and other journalists organized large datasets from the information released by Wikileaks, and even published them on Google Docs so they would be easy to share. "What if we put [the data] out there," he asked, "and let people help us generate stories?" Rogers estimated that out of this experiment came 10 to 15 published Guardian stories, many of which he estimated their journalists, himself included, may not have caught on their own.
Rogers' talk fit well into one of Day 2's more prominent themes -- data democratization -- both in that he and the Guardian were embracing crowdsourcing as an element of journalistic methodology, and because the visualizations the website employs are so interactive. In fact, the stories they tell can be considered dialogues between author and audience. For example:
Clicking the image above will take you to a Guardian page where you can play a time-oriented Flash animation depicting all recorded IED explosions in Afghanistan from 2004 through 2009. You can explore the periods of heavy conflict, the lulls and the groups that suffered the most casualties in each explosion, among other things. It is a great example of finding a design to fit the data, rather than the other way around (something Jock Mackinlay talked about in his presentation on Wednesday).
The other keynotes were equally interesting and effective. Scott Yara demonstrated that data proliferation --personal data proliferation-- is changing the way we do business and the way we communicate with each other, calling it a "step-function change" on the level of the PC revolution. He related a story in which his wife emailed him a link to photonotice.com, depicting the following:
There is Yara running a red light, whistling all the while. The image on the right was a Quicktime video catching him in the act. Somehow, even traffic light images of ourselves manage to find us.
Rounding out the keynotes were Barry Devlin of 9sight Consulting and The Heat Death of the Data Warehouse (video); a fantastic panel discussion on Posthumans, Big Data and New Interfaces (video) with Toby Segaran of Google, Amber Case of Geoloqi and Bradford Cross of Flightcaster, moderated by Alistair Croll of Bitcurrent; and, perhaps the best way to end by showing what data can do, a rousing talk by Carol McCall of Tenzing Health, who asked Can Big Data Fix Healthcare? (video). [SPOILER: The answer is a very interesting and qualified, "yes." She shows her work.] The big takeaway from her talk is that data can be used to identify early warning signs of oncoming illness, or to identify those at risk for particular diseases or disorders, thereby allowing for preventative measures. "We didn't find the answer," McCall told us, "we created a way to begin."
The first session I saw after the keynotes was on Data Journalism: Applied Interfaces with Rogers of the Guardian, Marshall Kirkpatrick of ReadWriteWeb and Jer Thorpe of the New York Times. The short talks from Rogers and Kirkpatrick were fascinating in their own right, both demonstrated well the uses of data in finding and telling journalistic stories. But it was Thorpe's visuals that really defined this session.
[Image from the New York Times]
This is a application code-named "Cascade" that the Times is internally using to visualize the way a story is shared through social space. Users can watch a story grow legs as it is shared and reshared on sites like Twitter. You can watch an amateur video shot at the presentation that gives an idea of what the program looks like in use here. Unfortunately, that video does not do the visualization justice. It is a powerful and beautiful application, really quite a treat to see.
Of course it would not be a day at Strata without some fascinating science behind how humans perceive things and how to harness those abilities. This was the talk by Dr. Creve Maples of Event Horizon, titled Beyond Visualization: Productivity, Complexity and Information Overload (though he preferred his original title, "Interconnections to the Mind"). Maples focused on our interaction with information, how we receive it and what we do with it. He looked at the total amount of relevant knowledge versus the amount that a person can keep up with, and how we might minimize the difference between the two--that "information anxiety" that populates what he called the "black hole between data and knowledge." He showed 3- and 4-dimensional models that were designed to utilize as much of the human brain's perceptive power as possible. And much as his talk revered the progress of computer automation, he emphasized the unique abilities of the human mind.
This is an image from Foldit, a project based at the University of Washington that relies upon user participation to contribute to scientific research. It is a game where humans compete against computers to recognize patterns in various protein folds, like HIV, Alzheimer's and cancer. The idea is to find out if humans' ability to detect patterns is superior to that of the computers and, if so, use that to improve the computers' models. When the game was first introduced, Maples told us, the results of the first head-to-head challenge were inspiring: humans beat the computer 5 out of 10 times, 3 tied, and 2 people did not answer; but in those last 2 instances, the computer's answers were wrong.
There was a certain irony in attending Maples' session. It was a talk about converting as much information to knowledge as possible, and yet it felt impossible to absorb all of the information he was giving us; it was like standing in a wind tunnel. Nearing the end of the third day, my brain felt full.
Thankfully, J.J. Toothman, a Web Strategist at the NASA Ames Research Center, anticipated that, and had just what we needed: beautiful pictures. Realizing that the conference itself was rounding out, Toothman told us he would give us a break and just show us "some really cool stuff." And he did not disappoint.
In a sense his presentation was a proper bookend to Naomi Robbins' talk, which minimized beauty in favor of message. Toothman did the opposite; message was there but the focus was aesthetics. Below are some of the more stunning projects he displayed.
[You can see Toothman's slideshow and find links to other data art on his website, here.]
A visual representation of character actions and flashbacks from the film "Adaptation." The link also includes images of the drawing process, which I highly recommend.
Twistori aggregates and anonymizes tweets that include the phrases "I love," "I hate," "I think," "I believe," "I feel," and "I wish."
Camille Utterback's Untitled 5 is an interactive painting that responds to motion. As her description puts it, "The existence, positions, and behaviors of various parts of the projected image depend entirely on people's presence and movement in the exhibit area."
A project out of MIT and on exhibit at the Museum of Modern Art, New York Talk Exchange demonstrates New York's connections to the rest of the world by illustrating long distance telephone and IP data flowing into and out of the five boroughs.
A Day in the Life of the MBTA uses Boston subway data to visualize subway activity. This particular image reflects all stations over the course of a 24-hour period.
Over the course of the conference, I was able to track down a few presenters and ask them a couple of questions. Tomorrow's final Strata 2011 post will include some varied ideas on Big Data, citizenship and the conference itself.
This post was written by Collin Sullivan. He is a research analyst for The Sentinel Project for Genocide Prevention, where data collection, analysis and visualization are being used to design an Early Warning System (EWS) to detect and prevent genocide. Collin lives in San Francisco. You can reach him at collin [at] thesentinelproject [dot] org and follow him on Twitter at @inciteinsight.