From smartphones and continuous data comes the social MRI

Dr. Nadav Aharony used phone sensors to explore personal behaviors and community trends.

It’s clear at this point that the smartphone revolution has very little to do with the phone function in these devices. Rather, it’s the unique mix of sensors, always-on connectivity and mass consumer adoption that’s shaping business and culture.

Dr. Nadav Aharony (@nadavaha) tapped into this mix when he was working on a “social MRI” study in MIT’s Media Lab. Aharony, who recently joined us as part of our ongoing foo interview series, described his vision of the social MRI:

“If you think about it, the three things you take with you when you go out of your home are your keys, your wallet and your phone, so our phones are always with us. In aggregate, we can use the phones in many people’s pockets as a virtual imaging chamber. So, one aspect of the social MRI is this virtual imaging chamber that is collecting tens or hundreds of signals at the same time from members of the community.” [Discussed at 1:16]

Aharony’s work focused on 150 participants (about 75 families) that were given phones for 15 months. During that time, more than one million hours of “continuous sensing data” was gathered with the participants’ consent. The data was acquired and scrubbed under MIT’s ethics guidelines, and for extra measure, Aharony included his own data in the dataset.

Collecting the data was just the beginning. Parsing that information and creating experiments based on emerging signals is where the applications of a social MRI became significant.

For example:

  • The phones can reveal the many online and offline networks people belong to via examination of online use, call logs, and even face-to-face interactions picked up by the phones’ built-in sensors. “You can construct multiple levels of networks for the same people and actually see how information flows, how friendships are made, how decisions are made in that community,” Aharony said. [Discussed at 2:30.]
  • The study revealed and filled participants’ memory gaps by combining automatic phone-generated data and manual surveys. “It’s interesting that people’s perception is not as effective as what you sense from the phone,” Aharony said. “If you ask people who they met this week, they probably won’t remember because our memories are a bit skewed. But the phone can remember who they saw around them during that week very accurately.” [Discussed at 4:56.]
  • And here’s an experiment marketers dream about: Aharony and his team wanted to see if there’s a relationship between social networks and app installations. Turns out there is, but it isn’t what you’d expect. There was no correlation between the apps people install on their phones and their self-reported networks. But there was a high correlation between the apps a person owns and the apps installed by people they’re around a lot (co-workers, parents gathering at a park while watching their kids, etc.). Proximity was the key. “It could be that 10 strangers might influence me more than my best friend,” Aharony explained. [Discussed at 5:16.]

The implications of Aharony’s study became clear to me when he drew a comparison to the long-running Framingham Heart Study, which surveys and examines participants every two years. “Think if you did this for 50 years with continuous data collection,” Aharony said. “Think about if you looked at who they spent time with, who they eat with, who they go out with, how these change over time, how your life changes when you have a baby … You can actually capture that.” [Discussed at 6:40.]

You know you’re on to something when you can make a legitimate comparison to one of the most important studies ever conducted. Yet, Aharony isn’t boasting, nor does he see his methodology as the end-all be-all for this type of data analysis. His enthusiasm and belief in the potential for this data is evident in the video interview — he zips through applications in research, medicine, personal health, city planning, disaster response and other domains.

What Aharony wants to do is create a framework based on the social MRI experiment that others can tap for their own needs. Toward that end, he’s teamed up with fellow MIT Media Lab members Alan Gardner and Cody Sumter to launch a startup called Behavio, which is producing an open source Android platform and toolkit that lets anyone track their own behaviors and potentially explore community-wide trends. “We want to build a platform to empower anyone to tap into, and take control of, their own data, and give them tools that help them use it to make their lives better and more productive,” Aharony wrote in a follow-up interview via email.

Behavio isn’t some side project that’s big on ideas but short on resources. It was recently awarded $355,000 from the Knight News Challenge.

Why this matters

Aharony and I chatted back in June, and since then, there have been two takeaways from the interview that keep bouncing around my head.

First, think about what it would take to conduct a study like this without smartphones. You’d have to goad participants into wearing sensors or keeping detailed journals or submitting to 24-hour surveillance. And even then, you still wouldn’t be able to gather the sheer amount of data that flows through a smartphone. Aharony stressed this point in our follow-up email:

“It is also important to note that this is not purely about the time collected; it is also about the number of signals, or dimensions, that were collected. It is one thing to collect a single signal — say, location — over those one million hours, but we were able to collect several dozen different signals for that same duration.”

A project like Aharony’s — something that harnesses this massive network — makes me see that I’ve been underestimating the mobile revolution. I don’t think I’m alone in that underestimation, either. Sometimes the mobile shift is simply too big to fully comprehend.

Second, smartphone adoption has created an ever-growing net of communication and the potential for a database that could eclipse all other databases. If the social MRI study, Behavio and other projects can responsibly capture and tap this data for the greater good (and economic gain … let’s not forget that part), we’ll shorten the distance between questions and answers.

We’ve already reached this point in a general-knowledge sense: Where’s the best local pizza place? Check Yelp. What’s the worldwide chicken population? Visit Wikipedia. How do I get from here to there? Fire up Google Maps. But if we have the ability to safely and easily acquire data at the personal, group and community levels, the questions we can answer become much more targeted and important: How do my current sleep patterns compare to a year ago? What factors influence my community’s decisions? Maybe even: Where can I get the help I need right now? All of this feels big.

Other things from the interview …

At the end of the interview, I asked Aharony about the people and projects he’s following. His interest in democratization and personal empowerment was evident in his choices. He pointed to the Quantified Self community — a movement Behavio clearly fits within — and the Open Source Satellite Initiative, which is looking to kick-start a wave of DIY satellites.

You can see the full interview from Foo Camp in the following video:

Associated photo on home and category pages: Big MRI by Muffet, on Flickr

Related:

tags: , , , , , , ,