Angela E.B. Stewart
Learning Scientist, Educational Technologist, Consultant
angelas [at] pitt [dot] edu
Current Projects
Girls as Technology Creators in CS Ed
While computing education programs abound, they continue to marginalize girls and people of color by centering dominant (white and male) perspectives. In my work, I explore how computing education can be justice-focused, and shift power towards those historically excluded in computing. For example, I leverage multimodal analytics to understand the multiplicitous expressions of engagement in diverse learners. Additionally, my work explores how to design culturally-responsive computing education environments that position girls as creators of technology, and encourage their technosocial change agency.
Collaborators: Amy Ogan, Erin Walker, Kimberly Scott, Tara Nkrumah, Leshell Hatley
Robots created by learners in a culturally-responsive computing camp for Black girls.
Classroom Data for Teacher Agency and Professional Development
Classroom discussions are an important pedagogical technique that promote learning and general reasoning competencies. Teachers are crucial to scaffolding these discussions. Yet they often struggle to promote rich discourse in their classrooms, particularly in science contexts. My work addresses this through a data visualization approach. I study how teachers use data in their classroom to support their agency in shaping discussions. I also engage teachers in participatory design, with the goal of creating data visualization tools that allow them to improve their pedagogical practice.
Collaborators: Amy Ogan, Sherice Clarke, John Zimmerman, Prasenjit Mitra, Ung-Sang Lee, Tricia Ngoon, Katherine Dennis
Teacher-designed visualizations of classroom data.
Completed Projects
Intelligent Systems for Collaborative Problem Solving
Collaborative problem solving is ubiquitous in everyday life, be it in education, workplace, or social scenarios. However, research has shown that people lack proficiency in CPS skills and teams often fail to perform as well as they theoretically could. To address this, I build fully-automated machine learning models of collaborative problem solving from multimodal signals, such as language, facial expressions, and acoustics. Taking a human-centered design approach, I embed these models into intelligent systems that provide dynamic feedback to teammates on their collaborative problem solving skills.
Collaborators: Sidney D'Mello, Valerie Shute, Nicholas Duran, Chen Sun
Graphic depicting the architecture of an intelligent system for collaborative problem solving (CPS) feedback.