I am a sucker for studies about the power of data and analytics to draw conclusions that can help us better understand the connections we form both online and offline, so I was immediately drawn to new research out of UNC about a new mathematical model for manipulating large data sets. The study’s lead author Peter J. Mucha, an associate professor of mathematics, said the “community structure” research he and his group performed is “connecting the dots on steroids.”
Mucha and his colleagues were able to account for shifting relationships between nodes of a network—say a relationship ending or friendships forming—with their advanced modeling techniques. The group analyzed data from Senate roll call votes and anonymous Facebook data to identify where communities formed and how they interacted.
The research should eventually allow user behavioral data to factor into the understanding of community action. The ability to measure how an individual user influences another—and the ability to quantify the force and duration of that influence—would be a powerful thing to know for researchers and marketers alike.
For instance, with the political data, the researchers were able to identify communities based on sources of donations and the regions they represented—among other traits. Mucha plans to use the technique to analyze the spread of disease next.
If these network modeling tools could be used to understand the reason and forces behind content “going viral” online, brands and businesses would have incredibly useful data about how valuable a customer advocate is—and how damaging a scorned individual customer could be to its business as a whole.
We’ll continue to monitor advances in understanding social data, because as these methods are refined and improved, brilliant conclusions and insights will certainly be produced. It has been impossible to date to accurately measure the spread of information and ideas, but Mucha’s research is the next step toward visualizing and understanding how we influence others around us.
