Vetria Byrd, Associate Professor of Computer Graphics Technology at Purdue University, will present "Building Data Visualization Capacity: A Bridge to Data Science and Data-Driven Decision Making."
Abstract: It is imperative that all persons have basic data visualization knowledge. Despite its growing need, there remains a disconnect regarding what to teach and how to teach data visualization to diverse audiences. In this talk, I will present recent and current research strands for building data visualization capacity as a bridge to data science that promotes data literacy, enhances information delivery, and builds knowledge and understanding to support data-driven decision making.
My research explores data visualization from two lenses: pedagogy and visualization of disparate heterogenous data. In this talk I will define “data visualization capacity,” describe a method for teaching and learning data visualization (Byrd & Dwenger, 2021) and share results from The Byrd Data Vis Data Science Lab’s research of using data visualization as a conduit for more advanced knowledge of data science and AI. I will present recent and current research results using Systemic Lupus Erythematosus (aka lupus) related data to model the complex system of disparate data. The Byrd Data Vis Data Science Lab has explored two of many research paths in the study of lupus. Our first preliminary study used generative adversarial networks (GAN) for lupus diagnostics. The principal methods of this work were based on artificial intelligence, machine learning, and medical imaging (Periasamy & Byrd, 2019). In this work, facial images with a lupus related skin rash, which typically appears on the face/cheeks called a “butterfly rash”, were used to train a machine learning algorithm to detect the presence of lupus and rule out other symptoms often misdiagnosed as lupus. Our second preliminary study focused on musculoskeletal manifestations of lupus using Literature Based Discovery (LBD) (Jaiyeoba & Byrd, 2022). Results highlighted the different treatment classifications in which treatment research can be explored. The complexity of lupus necessitates multi-level tools and resources to inform literacy about the disease. Our work explores diverse approaches and various perspectives in an attempt to gain a holistic view of the disease.
Lastly, generalizing insights from our previous studies, future directions include developing tools and techniques that will enable multi-level approaches (Data Visualization, Visual Analytics), computational techniques (Machine Learning, Data Science) to address and visualize social determinants affecting health, early diagnosis, treatment, diagnosis and access to quality care. Outcomes of this research will be useful for advancing basic and translational research on health disparities, enable aggregation of data and identification of potential research areas for analysis.
Bio: Vetria Byrd is an Associate Professor of Computer Graphics Technology at Purdue University. Dr. Byrd is a computer scientist and biomedical engineer by training and a visualization scientist by experience who uses data visualization, data science and computational methods to make sense of data. Dr. Byrd’s interdisciplinary background uniquely positions her to do research that focuses on exploring large scale data from multiple sources and visualizing complex data relationships in a format that informs but does not overwhelm the viewer. Dr. Byrd is the faculty lead for the campus-wide Data Mine Data Visualization Living Learning Community at Purdue. The Byrd Data Vis Data Science lab develops methods and tools to promote progressive learning paths to inform what and how to teach data visualization. Dr. Byrd is a member of the NSF I-GUIDE Education and Workforce Development Team where she is contributing curriculum content for data and geo-visualization. Dr. Byrd is the founder and organizer of Broadening Participation in Visualization (BPViz) workshops designed to broaden participation of women and underrepresented groups in STEM through visualization. She has given numerous national and international talks, data visualization trainings, BPViz workshops and outreach activities funded by NSF.
To attend virtually, email Christine Hopper for the Zoom link.