Using big data can determine societal connections, says renowned MIT computer scientist

<p>&nbsp;</p>
<p>Alex Pentland has recently been working on the Open Algorithms Project, which aims to obtain data without invading privacy.</p>

 

Alex Pentland has recently been working on the Open Algorithms Project, which aims to obtain data without invading privacy.


Alex Pentland, a professor at the Massachusetts Institute of Technology (MIT) and one of the world’s “most powerful data scientists,” according to Forbes, first became involved with phones even before they were invented. 

Pentland, who lectured at the Edward J. Bloustein School of Planning and Policy yesterday evening, said in the 1990s it was evident that computers were getting smaller and wireless. In what he called the “Pentland Project” at MIT, Pentland and his students built miniature computers powered by motorcycle batteries to simulate what it would be like to have a phone. 

“We learned a bunch of things,” he said. “The big thing was that we had millisecond-by-millisecond data about where people were, who they were talking to, what they were doing, what they were interested in … and that was completely revolutionary.” 

Before phones, social scientists would use survey data, observations and laboratory experiments to learn about human behavior. Using data from phones was significant because many of society’s actions were based on averages, which was difficult to determine from previous methods of data collection, Pentland said.

Today’s data also reveals that prior ideas that markets and globalization determined the best outcome of a society may be wrong. Instead, referencing Karl Marx, Pentland said it was the more minute interactions between people that mattered. 

In an example from the MIT Media Lab, Pentland referenced his research that won the Fragile Families Contest, which used data to predict which children would be at risk based on their circumstances. He found that though the United States government spends tens of thousands of dollars observing psychology, social work and sociology, these numbers added up to very little. 

“You could not do a good job predicting which kids were at risk from all of the money, measurement and science that went into it,” Pentland said. “However, if you look at neighborhood conditions, you could do a better job of predicting the average risk of kids.”

This showed that the “social fabric,” or the relationships between people, was possibly more important than data obtained from census surveys.

In order to balance between the possible invasion of privacy and oppression that may result from data collection, Pentland said one of the main principles that should be followed is to never share data unless absolutely necessary, which can be established by a query-and-response system. 

An example of this set-up is the Open Algorithms (OPAL) Project, which keeps data encrypted and allows only screened algorithms to obtain data about telecommunication, finance, health and social factors in a society. 

“It doesn’t ever query about individual things. It queries about aggregates, so (for example) how many people are in this census zone in the morning … this turns out to be a question that’s quite safe to ask and answer. It’s very hard to relate that to an individual,” Pentland said.

Using this OPAL Project, Pentland has made maps on various conditions in a society that reflect factors such as poverty, disease propagation, inequality and sustainability without endangering individual rights. For example, when applying this system to Senegal, it was found that moving the bus systems slightly would increase the efficiency of public transportation by 10%. 

This method has also been a way to form conclusions on the “aggregate flows of people,” Pentland said. On a local scale, this means using phone data to track where people work and shop. On a real-life scale, this system was used for tourism planning in Andorra, Spain. Data from people's phones reveals which areas have the most congestion, how many new tourists were in a region and even the wealth of the phone user. 

Another critical finding from aggregated data was that local choices affected the overall segregation of society. Studies revealed that people tended to go toward areas where there were people with similar backgrounds to them, which bred further segregation in mobility and even communication online. 

Pentland also warned about the ease of mobile phones, which leads to the majority of resources going to a singular provider, an example being how most people use select social media applications or shop from a limited number of websites. 

“The online environment is inherently promoting monopolies because transaction costs are so low and uniform,” he said. “We’ve built our society where longer term sustainability … is taxed locally. If the local transactions go away … then there’s no money left for investment in education, infrastructure, etc.”

Finally, in a call to action, Pentland emphasized the importance of the collective in changing the world. While one person downloading their digital record may not result in change, thousands of people using the data can compare their results to analyze larger trends.

“And if you have enough people doing that, you can go to Facebook and change their attitude,” he said.


Comments powered by Disqus

Please note All comments are eligible for publication in The Daily Targum.

Support Independent Student Journalism

Your donation helps support independent student journalists of all backgrounds research and cover issues that are important to the entire Rutgers community. All donations are tax deductible.