Assistant Professor Jana Diesner will speak at two upcoming conferences on the topic of collection and use of digital social trace data. Her talks will address current issues in this research field, including privacy, ethics and regulations, and methodological issues related to data accuracy as well as considering the content of text data for advancing social network theory.
Diesner and Julian Chin (MS '12), a research assistant in the GSLIS Center for Digital Inclusion, will speak at a workshop on human-centered data science at the nineteenth annual Conference on Computer-Supported Cooperative Work and Social Computing (CSCW). Hosted by the Association for Computing Machinery, CSCW will be held in San Francisco on February 27 - March 2. Diesner and Chin will present their paper, "Seeing the forest for the trees: considering applicable types of regulations for the responsible collection and analysis of human-centered data," at the workshop on February 28.
Abstract: In human-centered data science, people collect and analyze data about other people in order to gain knowledge and build methods and tools, among other benefits. This process is regulated by multiple sets of explicit and implicit norms and rules, including personal ethics, IRBs, privacy and security regulations, copyright, and terms of service. It is no trivial task for researchers to keep track of all applicable rules, identifying their meaning and compatibility, and practically implementing them. We review these types of regulations and outline why implementing them can be challenging. We conclude that educational offerings and institutionalized processes need to be developed and implemented so that scholars can gain the awareness, knowledge, and skills that are essential for responsibly collecting and analyzing digital social trace data. We also argue that scholars from this field need to be active participants in the public discourse and policymaking on this topic since they can contribute domain expertise and methodological and technical insights.
Diesner also will deliver a keynote address titled, “Rich and Reliable Signals: Making Responsible Choices for Working with Social Interaction Data,” at the General Online Research 2016 conference, which will be held on March 2-4 at Dresden University of Applied Sciences in Germany.
Abstract: Collecting and analyzing digital social trace data involves plenty of small choices with potentially big impact on research outcomes and derived implications and decisions. These choices refer to the recording and representation of data and settings for analysis methods and tools. While powerful computing and analytics solutions have helped us to scale up the overall research process, we have a poor understanding of the impact of these decisions on our findings, and insufficient best practices for documenting and communicating them. To address this gap, I will present our work on identifying the impact of accuracy issues with social interaction data on network analysis results.
Another often overlooked feature of digital trace data is the joint availability of information about social interactions and associated natural language use. I will present on our research on using text mining techniques to enhance network data with the ultimate goal of testing long-established network theories in unprecedented ways. I give an example of leveraging sentiment analysis to label graphs with valence values in order to enable triadic balance assessment of communication networks. Our method enables fast and systematic sign detection, eliminates the need for surveys or manual link labeling, and reduces issues with leveraging user-generated (meta)-data for this purpose.
Diesner’s research in human-centered data science and computational social science combines theories and methods from natural language processing, social network analysis, and machine learning. In her lab, researchers develop and advance computational solutions that help people measure and understand the interplay of information and socio-technical networks. They also bring these solutions into various application contexts, e.g. in the domain of impact assessment. For more information about Diesners’s work, visit http://people.lis.illinois.edu/~jdiesner/.
Diesner joined the GSLIS faculty in 2012 and is a 2015-2016 faculty fellow in the National Center for Supercomputing Applications at Illinois. She earned her PhD from the Computation, Organizations and Society Program at Carnegie Mellon University’s School of Computer Science.