Hi. My name is Jeff Rzeszotarski, and I am a PhD candidate in the Human-Computer Interaction Institute at Carnegie Mellon University. I am advised by Niki Kittur and have broad interests in data visualization, crowdsourcing, and social computing. My research focuses on helping both experts and everyday people make sense of complex data. I strongly believe that data big and small must be made accessible to as many people as possible, and I endeavor to encode that value into the systems I design and the research I pursue.

I have received a BA in computer science from Carleton College and a MS in human-computer interaction from Carnegie Mellon. My work has been featured publicly in venues such as TechCrunch and GigaOM, and I recently co-founded a startup, DataSquid, with the help of CMU Project Olympus and the AlphaLab Accelerator. I am a former Siebel Scholar, Carnegie Mellon Innovation Fellow, and Microsoft Graduate Research Fellow. In the past I have interned with Peng Dai and Ed Chi at Google and Merrie Morris at Microsoft Research.

Back at Carleton I developed a continuing interest in Japanese and Chinese art history, which has lead me to collect animation cels and ukiyo-e woodblock art. I am a strong proponent of undergraduate liberal arts programs. I make and sell 3d prints through my Shapeways store, and recently began throwing pottery. I love to travel, taking lots of photos along the way.

My name is pronounced "Jeff Rez-oh-tar’-ski" [dʒɛf ɹ̠ˤʷɛzoʊtɑɹ’skiː]. You can call me Jeff Rz (rez).
Feel free to email me at "jeff [dot] rzeszotarski [@t] gmail.com"


DataSquid is a web and mobile tool for analyzing complex multivariate data. Coming out of my research on new ways to interact with visualizations, the system helps users explore many dimensions of data at once with little or no training. I co-founded the company with Niki Kittur in June, 2014, and have had a lot of fun learning and building ever since then. Learn more at our barebones website and join our beta mailing list!

Research Projects

Naturalistic Visualizations

People are increasingly using mobile devices for everyday computing tasks, augmented by multi-touch interactions which break down the barrier between user and system through interactions that feel natural and match users' expectations. I develop data exploration techniques that use multi-touch and naturalistic interaction metaphors to closely match the sensemaking process users employ to make sense of complex data. By easily pivoting through dimensions, fluidly transitioning between views, and enabling direct manipulation of data, we can help users more deeply encode data relationships and identify outliers.

Crowd Observatory

Crowd labor markets such as Amazon Mechanical Turk, TaskRabbit and UpWork help employers to access large, instant-on pools of workers. However, the large scale of these markets limits their usefulness for complex or subjective tasks in which there is no single right answer. For example, there is no gold standard test question for tagging an image, and voting approaches don't work when workers write poems. In this project, my key insight is that the way workers work can often be as informative as workers' end products. Using a technique I call Task Fingerprinting, I apply machine learning to help stakeholders understand how and why their workflows are succeeding or failing.

Crowd Workflows

Crowd labor markets also hold the risk of overworking or fatiguing workers who must rapidly and accurately complete large amounts of tasks in short amounts of time. This poses risks to workers' health and wellbeing as well as task organizers' workflows. I explore introducing short, fun breaks into workflows to entertain crowdworkers as well as investigating how we might build systems that deliver the right task for the right worker at the right time, giving proper renumeration for effort and skill.

Collaborative Knowledge

Crowds of volunteers communicate and collaborate to build massive scale projects such as open source software, encyclopedias, and discussion forums. However, the very success of these systems generates with it a tremendous amount of historical data that pose a serious barrier to new contributors. For instance, past discussions and contributions to the Wikipedia article on Abortion amount to over 20 copies of Pride and Prejudice in length. Yet, in order to make successful contributions, newcomers must understand this information. I design novel visualizations powered by machine learning models that transform this historical data from a barrier into a beneficial resource for newcomers.


Asking public questions online is a complex decision for users, demanding a balance of social capital exchange and self-presentation maintenance. For example, a person may want to get health advice, but might avoid asking because they don't want to share their medical condition, or because they have recently asked their friends many other questions. Designers of question-asking services want to reduce these barriers, but lack a model of how users make this decision. I investigate the decisionmaking process people use on social networking services as they disclose and ask favors of others.


For a full list of publications and other professional details, please see my most recent CV.

Refereed Papers

Lasecki, W. S., Rzeszotarski, J. M., Marcus, A., & Bigham, J. P. (2015) The Effects of Sequence and Delay on Crowd Work. SIGCHI.

Dai, P., Rzeszotarski, J. M., Paritosh, P., & Chi, E. (2015) And Now for Something Completely Different: Improving Crowdsourcing Workflows with Micro-Diversions. CSCW.

Rzeszotarski, J. M. & Kittur, A. (2014) Kinetica: Naturalistic Multi-touch Data Visualization. SIGCHI. Best Paper Honorable Mention

Rzeszotarski, J. M. & Ringel Morris, M. (2014) Estimating the Social Costs of Friendsourcing. SIGCHI. Best Paper

Rzeszotarski, J. M., Spiro, E. S., Matias, J. N., Monroy-Hernández, A., Ringel Morris, M. (2014) Is Anyone Out There? Unpacking Q&A Hashtags on Twitter. SIGCHI.

Rzeszotarski, J. M. & Kittur, A. (2012) CrowdScape: Interactively Visualizing User Behavior and Output. UIST. Best Paper

Rzeszotarski, J. M. & Kittur, A. (2012) Learning from history: Predicting reverted work at the word level in Wikipedia. CSCW.

Rzeszotarski, J. M. & Kittur, A. (2011) Instrumenting the crowd: Using implicit behavioral measures to predict task performance. UIST.

Gross, D., Atlas, R., Rzeszotarski, J. M., Turetsky, E., Christensen, J., Benzaid, S., Olson, J. T., Steinberg, L., Sulman, J., Ritz, A., Anderson, B., Nelson, C., Musicant, D., Chen L., Snyder, D. C., Schauer, J. (2010) Environmental chemistry through intelligent atmospheric data analysis. Environmental & Modelling Software. 25, 6, 760-769.

Refereed Workshop Papers, Posters, & Presentations

Rzeszotarski, J. M. & Kittur, A. (2013) TouchViz: (Multi)Touching Multivariate Data. Interactivity Demo. SIGCHI.

Rzeszotarski, J. M. & Kittur, A. (2013) TouchViz: (Multi)Touching Multivariate Data. Work in Progress Poster. SIGCHI.

Rzeszotarski, J. M. (2011) Worker Collaboration in Crowdsourcing Markets. Workshop on Crowdsourcing and Human Computation: Systems, Studies and Platforms. SIGCHI.


Methods and Software for Visualizing Data By Applying Physics-Based Tools To Data Objectifications (Filed 2015)

System for Interactively Visualizing and Evaluating User Behavior and Output (Filed 2013)

System and Method of Using Task Fingerprinting to Predict Task Performance (Filed 2013)