andreavaccari.com

I just got back from San Diego, where I attended the ETech 2008 and presented the brand new New York Talk Exchange project, and I would like to offer my insider round up of the best of the best of the conference.

This year the focus was on “the brand new tech that’s tweaking how we are seen as individuals, how we choose to channel and divert our energy and attention, and what influences our perspective on the world around us.” In other words the eye was on the alpha geeks, as Tim O’Reilly calls them, who are working on ideas that are still in a early development phase, far from the publicity of the mainstream media and blogs but that will truly change our lives when they (or we) will hit maturity.

Day 1.

Tom Carden of stamen design gave a thorough introduction to data visualization. One by one, he presented different projects by stamen — Trulia Hindsight, Oakland Crimespotting, Backchannel, Digg Arc, Digg Stack — and others — Chris Jordan, the Dopplr Blog, the Graphviz and TouchGraph libraries, and geobloggers — to discuss what is data visualization and what are the best practices when starting to work on a new dataset. In particular, he suggested to try to visualize as much as possible of the the raw dataset, without filtering out any information. He also suggested that it is important to visualize the implicit information that is hidden in the data, and that it is vitally important to allow the user to interact with the data adding sliders whenever it is possible.

The evening events included a quite boring Facebook application session, where 6 facebook apps and 4 opensocial apps where presented. After that it was time for a great Tim O’Reilly to kick start the conference. His 30-minute speech was truly inspiring. After that the Ignite ETech started: 8 presenters, 5 minutes and 15 slides each, to introduce their works and their ideas. Among them my favourites where Saul Griffith, Matt Webb, and Scotto Moore.

Day 2.

During the morning keynote Eric Rodenbeck, creative director of again stamen design, argued that data visualization can be no longer considered just a tool for scientists and analyst who needs to explore large sets of data. In fact, data visualization is becoming a medium on its own, and as such it has a profound expressive potential. I believe this is a crucial issue: until now data viz has been used as a one-way tool to deploy pre-digested and pre-formatted information, and the user had no other choice than to read what the creator of the visualization decided to show. Now, however, as more people become aware of the unique potential of data visualization, the flow does not have to be one-way any longer, and even passionate users can contribute to capture the information hidden in the data with the help of social data visualizations.

Peter Norvig of Google presented different data-driven models for image searching and language translations. From finding a meaningful picture of a Starbucks coffee cup, to translating from Chinese to English, he explained that it is often easier to leverage the huge amount of data available on the Internet rather then foolishly try to find the perfect algorithm. For example, extracting different features from images tagged “starbucks”, clustering them together, and selecting the centre of their proximity graph will generate the desired result, without having to build complex image analysis algorithms. In the same way, selecting the correct semantic connotation of a Chinese word can be quite difficult if we try to actually understand the semantic meaning of the sentence, but it becomes quite easy when using a simple probabilistic model together with short text translations scraped from the Internet.

Paul Torrens described his work in modelling crowd behaviour. I was particularly fascinated by his research because it has strong ties with what we do at the Senseable City Lab. However, while we use a data-driven approach by collecting and analysing massive amount of data to understand how city dynamics are evolving in the urban environment, he instead starts from a agent-based descriptive model of how people move around the city, in normal conditions as well as in the case of an emergency, in order to develop tools that will help architects and urban planner in creating safer environments. Paul even run some motion capture experiments to calculate parameters as the average acceleration and deceleration of a person while walking, and added constraints to the behaviour of particular agents like a mother-son relationship where the two agents will do everything to find each other if separated.

Day 3.

Tom Coates of Yahoo! Brickhouse presented Fire Eagle. The aim of Fire Eagle is to enable location-based services that (a) manifest everywhere the network touches, (b) work seamlessly with each other, and (c) decouple the creation and use of data. Given different services that allow to get your position (like a GPS device, or the WiFi triangulation of the iPhone), and other services that can use such information to provide location-based services, Fire Eagle can work as a middle layer that allows you to define how the information flows from the former to the latter. In particular, you can define the resolution of your location (for example the exact latitude and longitude, or just the city, or the zip code), you can decide with whom to share your location, and more in general you can protect your privacy without having to renounce to such services.

Nathan Eagle, a former Media Lab Ph.D. student, talked about his ongoing project on reality mining. Between 2004 and 2005 he collected 400,000 hours of continuous human behaviour data related to the movements and social activity of hundreds of people (students, professors, and personnel) at MIT. Analysing location (derived through celltowers), proximity to others (derived through bluetooth), and communication patterns (derived through phone calls and text messages), he studied how different demographics behave differently in terms of entropy in time and space, how friendship can be inferred from proximity, how activities (sleeping, partying, …) can be inferred from the conjunction of location, proximity, and time of the day, and so fore. It is a truly wonderful initiative and he already achieved amazing results. I believe it would be even more wonderful to mine the same dataset to complement the analysis around the individuals with a study around the built environment.

Finally, from 8 to 9.30 pm, different artists presented their works during the Emergin Arts Fest. There I presented the brand new New York Talk Exchange project, currently on show at MoMA in NYC for the Design and the Elastic Mind exhibition. NYTE illustrates the global exchange of information in real time by visualizing volumes of long distance telephone and IP (Internet Protocol) data flowing between New York and cities around the world. Thanks to Brady and Kati who provided me two huge projections, attendees got immediately hypnotized by our visualizations and between one question and another, they took plenty of pictures and videos of the three animations.

Day 4.

During the closing keynote, Tim Ferriss talked about his new book: the 4-hour work week. Since you can’t ask him what he does, because he hates that, I will tell who he is: the most productive incarnation of anticonformism, the world champion of free-spirited individualism, the living example of lateral thinking and willful perseverance. In the last five years he won the Chinese kickboxing championship, became a record-holder in tango, and acted on a hit television series in Hong Kong. All of this, of course, while travelling around the world and controlling his firm from remote wireless spots. But it is not what he did that is most interesting, as much as how he did it. If you haven’t read the book, than go to the book store right now!