I am the Science Director at Graphika, Inc, a social media analytics startup based in New York.
My research focuses mainly on the intersection of social media and large-scale social phenomena. I am keenly interested in how large groups of people behave together, how ideas and movements spread from individual to individual, and how much we can deduce about a person's behavior from their digital footprint.
In addition to my research duties, I oversee many aspects of Graphika's technology stack, from automated discovery of key content for the platform to building the Graphika API and generating custom user reports.
Respondent-Driven Sampling - Testing Assumptions: Sampling with Replacement
Vladimir D. Barash ◦ Christopher J. Cameron ◦ Michael W. Spiller ◦ Douglas D. Heckathorn
Berkman Russia Paper Series
Karina Alexanyan ◦ Vladimir Barash ◦ Bruce Etling ◦ Robert Faris ◦ Urs Gasser ◦ John Kelly ◦ John Palfrey ◦ Hal Roberts
Tracking Critical-Mass Outbreaks in Social Contagions
PI: Michael Macy, Cornell University
Co-PI: Clay Fink, Johns Hopkins Applied Physics Laboratory ◦ John Kelly, Graphika, Inc. ◦ Vladimir Barash, Graphika, Inc.
Analyzing Social Media Networks with NodeXL:
Insights from a Connected World
Chapter 9: Twitter: Conversation, Entertainment, and Information, All in One Network! Marc Smith, Ben Shneiderman, Derek Hansen. Morgan Kaufmann Press, Waltham MA 2010
2010: Chapter 9: Twitter: Conversation, Entertainment, and Information, All in One Network! In Analyzing Social Media Networks with NodeXL: Insights from a Connected World. Marc Smith, Ben Shneiderman, Derek Hansen. Morgan Kaufmann Press, Waltham MA 2010
Classical Respondent-Driven Sampling (RDS) estimators are based on a Markov Process model in which sampling occurs with replacement.
We join recent work by Gile and Handcock in exploring the implications of the sampling-with-replacement assumption for bias of RDS estimators. We differ from previous studies in examining a wider range of sampling fractions and in using not only simulations but also formal proofs.
One key finding is that RDS estimates are surprisingly stable even in the presence of substantial sampling fractions.
Social media sites like Twitter enable users to engage in the spread of contagious phenomena: everything from information and rumors to social movements and virally marketed products. In particular, Twitter has been observed to function as a platform for political discourse, allowing political movements to spread their message and engage supporters, and also as a platform for information diffusion, allowing everyone from mass media to citizens to reach a wide audience with a critical piece of news. We leverage previous work to construct a system for classifying contagious phenomena based on the properties of their propagation dynamics, and apply our system to a dataset of news-related and political hashtags diffusing through the population of Russian users of Twitter. Our results show that news-related hashtags have distinctive propagation dynamics, but that political hashtags have many different dynamic signatures.
Who is Vlad?
I received my Ph.D. from Cornell University, where I studied Information Science and wrote my thesis on the flow of rumors and virally marketed products through social networks.
Although I was born in Russia, I have lived up and down the east coast of the US most of my life. These days I am anchored in the greater Boston area. Here, I pursue my passions of science, writing, and games alongside my partner and our little dog, whom we love very much.