Bowdoin College
  • RSS
  • Email
  • Twitter

Featured Project

Big Data, Social Networking Technology, and the Life Sciences

Project Overview

leverage big data-driven research methods to reveal new insights into the way life scientists use diverse social networking and social media technologies to network, collaborate, & maintain virtual scientific communities.

Introduction

Our lab is currently engaged in a project that explores the activity and cyber-infrastructure of Virtual Organization Breeding Environments  (VBEs), virtual spaces which are responsible for encouraging the formation of virtual organizations (VOs). Specifically, our research seeks to explore collaborative virtual organizations that form around the study of the life sciences. Our goals are to assess the ways that VOs are utilized in the life sciences and to understand their potential to foster trust, collaboration, and social cohesion between potential VO team members, crucial prerequisites for VO success. In this project, we seek to extend our observations of life science VOs into social networking and social media technologies.

Goals of the Project

Cast a wide net to capture and index a broad-spectrum sample of Twitter and social media activity within a scientific community.

Develop identification criteria for the discovery of collaborative communities, discussion of life sciences in social media, and key users within these communities by utilizing topic analysis, machine learning, social network analysis, and other advanced techniques.

Discover collaborative clusters using criteria and models developed in the previous phase. Create a map of a broad survey of science-based communities on Twitter and other selected social media.

Investigate and analyze cluster behavior through in-depth analysis of the activity and infrastructure that allows these communities to be successful or not.

Why us?

The Social Network Innovation Lab has developed expertise in the study and analysis of social networking and social media technology use in the life sciences. Over the last three years, the lab has been collecting relevant data for sociological analysis. The lab has engaged in a number of research projects using this data, including large-scale topic and sentiment classification projects, advanced behavioral analytics, social network analysis, major event and disasters patterns, and demographic explorations. Our work has involved the development of custom research methods built around advanced data mining, social network analysis, text analysis, and automated classification techniques.

Core Data and Goals

AWS_ac_ra_largescale_05
We propose to collect a diverse set of social media data relevant to life science communities. We anticipate to collect approximately 5-10 million social media objects a day (i.e. posts, comments, images). With this data, we will deploy a number of mapping and classification algorithms to identify scientific activity, discussions, and collaboration taking place within social media. After identifying where and how the clusters of activity operate, we will  focus in on these communities to capture a higher percentage of explicit  collaborative activity which will be subjected to qualitative  analysis in addition to quantitative network and sentiment analysis.

Goals

awsgraphic-2

With this project, we hope to bring the power of big-data analytics to the identification and study of collaborative communities in the life sciences. A model for the successful identification of collaborative activity taking place in social media would be highly valuable to the study of online social behavior in any field, but especially where the sciences are concerned. These identified VOs would allow for a comprehensive understanding of how social media is being used and is influencing modern scientific practice. Further, the methods would allow for in-depth comparative studies of individual VOs on social media and the possibility to determine which are most successful and why. Understanding the properties of successful VOs is of critical importance in the design of optimal and productive virtual scientific teams.