SocioPatterns

The SocioPatterns project was born out of an effort to obtain more and better data for use in research on the dynamic characteristics of complex networks, in particular social networks. Traditional methods, i.e., observation and self-reporting, simply do not allow for the kind of large-scale, unbiased and temporarily fine-grained data we were looking for. We started with an experiment in which we used the active-RFID badges developed by the OpenBeacon project to collect temporarily fine-grained data on the evolution of the social network in a medium-scale population. Based on the comparison of reception patterns of the beacon signal emitted by the badges, the system inferred proximity among the participants. However, mere proximity is not a good proxy of social contact. We therefore rewrote the firmware and collaborated with the nice people of Bitmanufaktur, the OpenBeacon developers, to produce tailored hardware so that we could directly sense face-to-face proximity by using a low-power peer-to-peer communication scheme. Some of the various badges produced during this process are shown in the following image. They measure 35 by 35 by ~7 mm.

We have since used the resulting SocioPatterns platform in more than ten experiments in different locations and contexts: various conferences in Italy, France, Germany and Greece, a three-month exhibition in Ireland, a primary school in France, and a children's hospital in Italy. While for some of these experiments we focused on pure data-collection, for others we were also interested in real-time social services made possible by the availability of real-time data on the social interactions among a group of people. One of the service oriented experiments led to the Live Social Semantics project.

In order to monitor the data collection system as well as to demonstrate it, we've created real-time visualization applications. This application, developed for the cross-platform Adobe AIR runtime, connects with a server from which it receives regular updates of the contact network. The following movie shows a time-lapsed replay of this visualization for some interesting segments from one day of one of the experiments.

Compare the above results with those of the first experiments shown in the next movie. While in the first experiment, shown below, all proximity relations are detected and thus shown, in the subsequent experiments, shown above, only face-to-face proximity relations are shown. The movie below clearly shows much more links that form a cliquish pattern, which is not what we were after.

Compare the above results with those of the first experiments shown in the next movie. While in the first experiment, shown below, all proximity relations are detected and thus shown, in the subsequent experiments, shown above, only face-to-face proximity relations are shown. The movie below clearly shows much more links that form a cliquish pattern, which is not what we were after.

The following figure shows how the visualization was publically displayed during some of the experiments.

Some of the deployments of the visualization during various experiments. Top-left: ESWC'09; top-right and bottom left: Infectious in Dublin; bottom-right: Epiwork08.

More information:

• SocioPatterns.org - the project site
  • Exposing contact patterns - about the very first experiment
  • New firmware for contact detection - about the firmware rewrite
  • Contact duration and Social interactions at a conference - some first analysis results
  • Contact Patterns part 2 - about the principal real-time visualization
  • SocioPatterns at Infectious in the Science Gallery, Dublin - about the 3-month deployment at an artscience exhibition in Ireland
• Live Social Semantics - a collaboration of the SocioPatterns and TAGora projects

In collaboration with:

Ciro Cattuto, Alain Barrat, Brita and Milosch Meriac, Jean-François Pinton, Prof. Alessandro Vespignani, Vittoria Colizza, Marco Quaggiotto, Lorenzo Isella, and Juliette Stehlé.

Publications

• What's in a crowd? Analysis of face-to-face behavioral networks. Lorenzo Isella, Juliette Stehlé, Alain Barrat, Ciro Cattuto, Jean-François Pinton, and Wouter Van den Broeck. Journal of Theoretical Biology. Volume 271, Issue 1, 21 February 2011, Pages 166-180.
• Wearable sensor networks for measuring face-to-face contact patterns in healthcare settings. Alain Barrat, Ciro Cattuto, Vittoria Colizza, Lorenzo Isella, Caterina Rizzo, Alberto E. Tozzi and Wouter Van den Broeck. In: Proceedings of the 3rd International ICST Conference on Electronic Healthcare for the 21st century (eHealth 2010). Casablanca, Morocco, December 2010. (Won conference best paper award)
• Dynamics of Person-to-Person Interactions from Distributed RFID Sensor Networks. Ciro Cattuto, Wouter Van den Broeck, Alain Barrat, Vittoria Colizza, Jean-François Pinton, and Alessandro Vespignani. PLoS ONE 5(7): e11596. doi:10.1371/journal.pone.0011596, July 2010.
• High resolution dynamical mapping of social interactions with active RFID. Alain Barrat, Ciro Cattuto, Vittoria Colizza, Jean-Francois Pinton, Wouter Van den Broeck, and Alessandro Vespignani. Informal publication, Nov 2008.
• See also Live Social Semantics publications.