SAEIDEH BAKHSHI | PhD student in Computer Science, Georgia Institute of Technology
I am currently a Computer Science PhD student at Georgia Tech, working under the supervision of Dr. Eric Gilbert. I am also a member of Comp.social lab. I received my Masters from Sharif University of Technology and Bachelors from Amirkabir University of Technology.
I am interested in understanding users and how they interact with online content. My goal is to use the findings of user behavior research to enable design of systems that enhance social relationships and improve user experience. I mostly use data mining approaches to understand and evaluate online technologies. My research is at the intersection of social computing, data mining and HCI. I interned at Yahoo Labs during Fall 2013 semester working with Dr. David A. Shamma.
I love to write code, drink good coffee and read good books.
I am planning to graduate in summer 2014 and seeking a full time research scientist position. Please contact me at saeideh (at) gatech (.) edu for my application materials.
|Faces and Photo Engagement|
|Title: Faces Engage Us: Photos with Faces Attract More Likes and Comments on Instagram
Abstract: Photos are becoming prominent means of communication online. Despite photos’ pervasive presence in social media and online world, we know little about how people interact and engage with their content. Understanding how photo content might signify engagement, can impact both science and design, influencing production and distribution. One common type of photo content that is shared on social media, is the photos of people. From studies of offline behavior, we know that human faces are powerful channels of non-verbal com- munication. In this paper, we study this behavioral phenomena online. We ask how presence of a face, it’s age and gender might impact social engagement on the photo. We use a corpus of 1M Instagram images and organize our study around two social engagement feedback factors, likes and comments. Our results show that photos with faces are 38% more likely to receive likes and 32% more likely to receive comments, even after controlling for social network reach and activity. We find, however, that the number of faces, their age and gender do not have an effect. This work presents the first results on how photos with human faces relate to engagement on large scale image sharing communities. In addition to contributing to the research around online user behavior, our findings offer a new line of future work using visual analysis.
Publication: Bakhshi, Shamma, Gilbert | CHI'14
|Factors affecting Restaurants' reviews|
|Title: Demographics, Weather and Online Participation: A Study of Restaurant Recommendations
Abstract: Online recommendation communities are valuable information sources that people contribute to, and often use to choose restaurants. How- ever, little is known about the dynamics behind participation in these online communities and how the recommendations in these communities are formed. In this work, we take a first look at online restaurant recommendation communities to study what endogenous (i.e., related to entities being reviewed) and exogenous factors influence people’s participation in the communities, and to what extent. We analyze an online community corpus of 840K restaurants and their 1.1M associated reviews from 2002 to 2011, spread across every U.S. state. We construct models for number of reviews and ratings by community members, based on several dimensions of endogenous and exogenous factors. We find that while endogenous factors such as restaurant attributes (e.g., meal, price, service) affect recommendations, surprisingly, exogenous factors such as demographics (e.g., neighborhood diversity, education) and weather (e.g., temperature, rain, snow, season) also exert a significant effect on reviews. We find that many of the effects in online communities can be explained using (offline) theories from experimental psychology. Our study provides a first insight into the dynamics of the highly popular online restaurant recommendation sites. It has implications for designing online recommendation sites, and in general, social media and online communities, to be more effective.
Publication: Bakhshi, Kanuparthy, Gilbert | WWW'14
|Statistical overview of Pinterest|
|Title: “I Need to Try This!”: A Statistical Overview of Pinterest
Abstract: Over the past decade, social network sites have become ubiq- uitous places for people to maintain relationships, as well as loci of intense research interest. Recently, a new site has exploded into prominence: Pinterest became the fastest social network to reach 10M users, growing 4000% in 2011 alone. While many Pinterest articles have appeared in the popular press, there has been little scholarly work so far. In this paper, we use a quantitative approach to study three research questions about the site. What drives activity on Pinterest? What role does gender play in the site’s social connections? And finally, what distinguishes Pinterest from existing networks, in particular Twitter? In short, we find that being female means more repins, but fewer followers, and that four verbs set Pinterest apart from Twitter: use, look, want and need. This work serves as an early snapshot of Pinterest that later work can leverage.
Publication: Gilbert, Bakhshi, Chang, Terveen | CHI 2013, pdf
|Role of Photo Content in Engaging Users|
|Research question: What is the role of content of the photo on it’s engagement value? What is the emotional value of each type of content?
|Filters and Photo Engagement|
|Research question: How image filters, as proxies to artistic visual transformation, may impact users’ implicit consumption and explicit action on content?
Data: 7.9M Flickr mobile photos
|Colors and Diffusion of Images|
|Research question: What is the relationship between dominant color of an image and it’s virality on social media? We know from previous research that colors affect behavior, is this phenomena observable online?
Data: 1M Pinterest pins and corresponding users, images corresponding to every pin
|Effect of media type on engagement and diffusion|
|Research question: What is the role of images, as communication content, in engaging users? How do they compare to the more traditional word-based content?
Data: 8K Facebook pages, 7M Facebook posts
Methods: Negative binomial regression, Confidence intervals
Tools: ggplot2, lots of R packages, Facebook Graph API
|Topic modeling on Social media|
|Research question: How does LDA perform on twitter (with social media language) compared to Wikipedia and New York Times articles?
Data: 1M tweets, 50K wikipedia articles, 5 years of NYT articles
Methods: LDA, Comparison of enthropies, Efficiency analysis
Tools: Mechanical Turk, LDA implementation
|Inferring brain network of depression|
|Research question: How are different areas of brain
(important in depression) connected to each other? What is the role of
SubCallosal Cingulate in this network?
Data: DTI brain images of 30 healthy subjects
Methods: Tractography, Network inference, Centrality analysis
Tools: FSL, Matlab, Brain connectivity Toolbox
Publication: Bakhshi, Gutman, Dovrolis | INCF 2011, link
|Inferring brain network of depression|
|Research question: What are the differences of designing a network optimally and evolutionary?
Data: Simulation results
Methods: Clustering, Network analysis
Tools: Java, Matlab, Brain connectivity Toolbox, gnuplot
Publication: Bakhshi, Dovrolis | Bionetics 2011 pdf pptx
Bakhshi, Dovrolis | IFIP Networking 2013 pdf