President's Dream Colloquium on Engaging Big Data

Watch the Recorded Web Stream

Recorded on March 1, 2016

About the Lecture

This lecture will introduce students to the sociotechnical shaping of 'big data.' The development of big data technologies and their adoption is both technical and social.

Similar to many innovative technologies, big data has being going through a phase of diffusion in the past couple years characterized by booster and skeptic discourses.

Data entrepreneurs and critics grapple over the technical and social stories about the benefits, challenges, and risks of big data. For example, Google Flu Trends has been the poster child for innovation, creativity, and public impact of big data. Boosters such as Mayer-Schonberger & Cukier trumpet GFT as a success story of analytic success and efficiency over traditional approaches and domain experts.

Some are skeptical of big data replacing domain expertise. In a recent paper in Science, David Lazar and colleagues challenge the hubris of big data by examining GFT's record of accurately predicting flu outbreaks. Their findings show that GFT did a pretty good job overall. However, when the algorithm was off, it was way off.

This points to the need for a meeting ground between the technical and the social. For big data approaches to be effective, data scientists need to understand users, various stakeholders, and cultural context.