Contractor Computes Solutions to Real-World Problems

By Alan Wong - Advances in computing and data technology have allowed social scientists to develop exciting new solutions to long-standing problems. At the Gerald P. Mohrmann Memorial Lecture on May 18, 2017, Noshir Contractor explored some of those solutions in a talk entitled "Leveraging Computational Social Science to Address Grand Societal Challenges".

Contractor, Jane S. & William J. White Professor of Behavioral Sciences at Northwestern University, studies social networks. In his lecture, drawing on his own research in the area of (social) networks, he presented three examples of how more data and increased computational power can motivate new theories and methods. He also demonstrated how social science theories can be applied to real-world problems.

First, Contractor discussed applying network theory to disaster relief efforts, using the example of Hurricane Katrina. Could FEMA have responded more effectively? One source of data which quickly becomes available in the wake of natural disasters is “situational reports”. Sitreps, as they are known, are unstructured text documents describing events or occurrences. 

Going through sitreps manually is very time-consuming. For one event, doing so might take half a year or more, including the training of many undergraduate students to code and input useful data contained in the reports. Automating this process with computerized text-analytic methods could, Contractor said, achieve the same results virtually in real time. While we cannot predict when an event like Hurricane Katrina may occur next, pre-emptively building tools based on network theory would enable us to respond faster and better when it does.

Building unbeatable teams

Next, Contractor showed how research into networks might help people to build better teams. Teams of highly skilled individuals can produce results which are beyond the capability of any individual member, especially as projects grow in complexity. As an example, Contractor mentioned David A. Ferrucci. In a New York Times article, Ferrucci—who led the team that created an A.I. system capable of defeating human opponents on Jeopardy!—describes the challenges of assembling teams for big projects, especially high-risk ones that demand top talent. How might the formation of teams like Ferrucci’s be studied?

One context in which teams are formed with clearly measurable outcomes is in Massive Multi-Player Online games. Many such games require players to form teams. Since success or failure is already quantified in game data—in metrics such as numbers of monsters killed, player deaths, and time to complete objectives—researchers can use that data to compare more diverse teams with more homophilic ones. 

Saving young lives

Finally, Contractor shared research on how network theory may be used to address global health problems, such as neonatal mortality. Many neonatal deaths are preventable, provided the correct treatment options are available. One solution found to be both very cheap and effective for combating infections leading to neonatal mortality is chlorohexidine, an antiseptic. Using network theory combined with computational methods, Contractor and his colleagues studied how this solution could be scaled up to save many more lives.

Learn more about Noshir Contractor.