I am a PhD Candidate at the School of Information & Computer Sciences at University of California Irvine. I use computational methods and large scale experiments to understand linguistic and social behavior in automated environments, ranging from autonomous vehicles to people’s social media news feeds.
My current work focuses on understanding the various factors that impact the quality of democratic discourse on the web (see my most recent work here).
More recently, I worked at Facebook (Menlo Park HQ) as a PhD Research Intern, analyzing large-scale user data – comments & interactions on Facebook’s Live.
Some of my latest work
Fostering Civil Discourse Online?
One year after the #MeToo movement, reports show that people are more divided on issues surrounding sexual harassment & assault, with the divide running primarily along partisan lines. Want to know why this might be the case? Through this work, I use computational methods (like TF-IDF, shown in the chart above) to demonstrate how topics surrounding an online social movement are framed and conversed across people who consume news sources that are politically distinct (alt-right, far-left, and mainstream). Click here for the full paper.
Do people become more privacy-sensitive on the web as economic living standards improve?
This work explores online privacy and security attitudes from 24,143 individuals across 24 countries with diverse economic living standards. By using k-mode clustering, we identified three distinct profiles based on similarity in Internet security and privacy attitudes measured by 83 items. By comparing the aggregated dissimilarity measures between each respondent and the centroid values of the three profiles at the country level, we assigned each country to their best-fitting privacy profile. We found significant differences in GDP per capita between profiles 1 (highest GDP) to 3 (lowest). People in profiles with higher GDP per capita have significantly greater privacy concerns in relation to information being monitored or bought and sold. These individuals are also more reluctant towards government surveillance of online communication as well as less likely to agree that governments should work with other public and private entities to develop online security laws. As economic living standards improve, the proportion of individuals increases in profile 1, decreases in profile 2, and most rapidly drops in profile. To the best of our knowledge, it is the first research that systematically examines country-level privacy in relation to a national economic variable using GDP per capita. Click here for the full paper.
This work was sponsored by Microsoft Research and NSF
How do you express issues related to low-income identities on social media when everyone in your network is rich?
In this work, my co-authors and I examine stigma related to class identity on social media using a class confessions page of an Ivy League University that we call EUCC – short for Elite University Class Confessions. EUCC is an online space that includes a Facebook page and a surrounding sociotechnical ecosystem. In this work, we discuss how EUCC’s design shapes the nature of user interactions around class stigma, and explore in depth how people experience stigma differently through the restorative properties of EUCC. Click here for the full paper.