Data Science Problem Solving
Mastery of data science problem-solving requires students to synthesize their declarative and procedural knowledge and skills in data science methods and procedures and ultimately develop “adaptive expertise” to solve various real-world problems. Those skills are high in demand; however, they are hard to teach. To provide students with opportunities for deliberate problem-solving practices, we are developing a repository of caselets (= bite-size case studies) that mimic the real-world data science problem-solving processes. This project was initially funded by the Simon Initiative at Carnegie Mellon and piloted on the Open Learning Initiative platform (https://oli.cmu.edu/). At UMBC, together with Dr. Shimei Pan, we are looking into how to weave ethical considerations into the data problem-solving skills training with a seed grant from the Academic Data Science Alliance. With fine-grained data to be collected from students while working through caselets, we are also exploring the cognitive and metacognitive processes of data science problem-solving, ultimately providing personalized and adaptive support to data science students at a large scale.
Colalboraor: Dr. Shimei Pan (UMBC IS)
Student Researchers:
- Maryam Alomair (UMBC IS Ph.D. student, Fall 2021 – present)
- Megha Singh (UMBC CSEE Master student, Fall 2023 – present
- Kiran Prabhu (UMBC IS Ph.D. student, Summer 2022)
- Lucky Verma (UMBC CSEE Master student, Spring & Summer 2022)
Funding sources:
ADSA/Career Development Network
DARPA AI Tools for Adult Learning Prize
Data Storytelling
Data storytelling has seen exponential growth in real-world demand in recent years. It is increasingly becoming a consensus that strong analytical and programming skills alone are insufficient to communicate the value of data effectively. These skills need to be complemented with data storytelling competence in constructing coherent and compelling stories or narratives for audiences of various types. Despite its high demand and essential role in determining the ultimate fate of data-driven projects, education practices have not provided methods to train these specific skills on a large scale. Moreover, the lack of relevant educational research will hinder us from delivering optimal training experiences to diverse learners on these critically needed skills. To scale up the training of proficient data storytellers with limited teaching resources, we need to invest in cross-disciplinary educational research to support scalable teaching and learning in data storytelling. This project will leverage convergent expertise from learning science, computer science, data science, communication science, and human-centered design to explore an intelligent coaching support platform enabled by computational models and tools with embedded formative assessment and scaffolding functions.
Collaborator: Dr. Jiaqi Gong (The University of Alabama) and Louise Yarnall (SRI International)
Student Researchers (UMBC Site)
- Yetunde Okueso (Ph.D Student in HCC program, Fall 2023 – present)
- Jennifer Posada (Master student in HCC program, Fall 2023- present)
Funding source
National Science Foundation RETTL Project (Study Studio: Coaching Data Storytelling at Scale)