Jessica Ann Roberts


My research focuses on human-data interactions. I am particularly interested in spatially referenced data (data maps) in informal learning contexts like museums. Because large data sets are increasingly available to anyone through online repositories, as are do-it-yourself Geographic Information System (GIS) and data visualization tools, we encounter data maps and representations more frequently than ever before. Yet, these representations only re-represent the raw data for us: they do not do the actual work of interpretation. I hope through my work to better understand how people learn using these tools and how we can help learners become active rather and passive consumers of data.

Current Work

I am currently a postdoctoral research fellow at Carnegie Mellon University's Human-Computer Interaction Institute. We are developing a teaching collection of aquatic macroinvertebrates to help water quality biomonitoring organizations in the eastern United States train their citizen science volunteers to identify specimens. By improving supports for trainers and learners, we hope to improve citizen scientists' accuracy, confidence, and engagement in macroinvertebrate identification.

Past Work

As a postdoctoral researcher at Northwestern University, I engaged in an interdisciplinary collaboration between Northwestern and the Field Museum of Natural History studying touchscreens installed in the new permanent exhibition the Cyrus Tang Hall of China. Through observations, video analysis, and visitor tracking, we are investigating how multiple interface design iterations affect visitor engagement with exhibit artifacts.

My graduate research involved CoCensus, an interactive census data map exhibit at the New York Hall of Science. This exhibit aims to support visitors in spatial, temporal, and quantitative reasoning about mapped census data by allowing them to self-select four categories of census data (heritage, household size, housing type, industry) and manipulate the aggregation level (census tract, borough, or city) and the decade of data (1990, 2000, and 2010) shown on a large shared display. Through these interactions and manipulations in the multi-user exhibit, visitors can engage in open-ended explorations and conversations about patterns. Research explored how interaction design facilitated visitors' learning talk as they made sense of the data together.

My dissertation research examined how competing interaction designs influence visitors' reasoning talk during group interactions. Through a 2x2 interaction design I examined the impacts off the means of control (full-body interactivity versus a handheld tablet controller) and the distribution of control (single input in which a control action affects all data simultaneously, and multi-input in which each user can manipulate his or her own data individually) on visitors' data talk and interactions in an in situ study. Findings from this study were reported in the International Journal of Computer Supported Collaborative Learning (DOI 10.1007/s11412-017-9262-x). I also investigated the spontaneous use of actor perspective taking (APT) during interactions and explored how APT was used by visitors to relate to the data. My methodology for measuring learning talk was awarded Best Paper at the 2017 International Conference of Computer Supported Collaborative Learning (CSCL).

For prior studies and analyses related to this project, including the pilot work for my dissertation study, please see my curriculum vitae page.