Getting Local with Data: Possibilities and Pitfalls Across Three US Place-based Data Science Research Projects

• In Event: Data Science Education: Re-Envisioning for 21st Century Teaching, Learning, and Research

Abstract

Place-based data science education can foster agentic, relevant learning opportunities with and about data (Rubel et al., 2017). Yet the practice complexities of designing and enacting data science with youth, educators, and community organizations—including local power dynamics (Van Wart et al, 2020), material aspects of storage, organization, and visualization (Barany et al., 2024), and facilitation - remain overlooked. Comparing three US data science education research projects, we ask: What possibilities and pitfalls emerge when designing and enacting local data science endeavors?

Data and place are socially constructed, data through a dance of material, disciplinary, and human agencies (Hardy et al, 2020) and place as space made meaningful by humans (Tuan, 1977). We conceptualize place-based data science as grounded in hyper-local knowledge from learners’ lives.

We present work from three place-based data science education projects: 8-10 year-old children and educators in Richmond, VA; 11- and 12-year-olds in Queens, NY; and middle and high school science teachers in the Great Smoky Mountains. Each multi-year research project designed curricula and collected meeting notes, field notes, and youth/educator-authored artifacts (guiding questions, data sets, visualizations). Using thematic analysis (Braun & Clarke, 2006), we identified shared possibilities and pitfalls across our projects, illustrating each with an example.

Our projects document the rich ways that learners use experiential knowledge to make sense of collected data. Learners’ applications of experiential knowledge often shift power dynamics, as when a youth in Queens, NY, contested a data display by asserting “All of my group played hole #9, which is more than 2 people!” (Figure 1, left). Learners also creatively express their local knowledge within data displays, as when a child studying squirrel populations in Richmond, VA used a reflective drawing to emphasize noise levels near the highway (a deterrent for nesting) to explain the low tally for Trip 2 (Figure 1, right).

Figure 1. Left - facilitator-generated graph of youth mini-golf data. Right - youth reflective drawing and class pictorial data.

One pitfall of place-based data science education concerns the questions driving data science inquiries: their authorship, feasibility, and relevance to learners’, educators’, and communities’ goals. In Queens, NY, we developed a Venn diagram activity to emphasize dimensions of driving questions (e.g., Possible, Data Driven, and Relevant), posing, placing, and discussing questions each day, though facilitators expressed uncertainty about if/how to correct learners’ question placement. A second, shared challenge is the time and resources needed to scaffold learners’ data explorations, especially in processing, organizing, and/or visualizing learner-collected data. In a place-based data science PD with Tennessee teachers, processing and cleaning the place-based data collected by one class of high school students required 15+ hours, raising questions about the sustainability of using rich but hard-to-process data sources. We must develop frameworks and systems to resolve these pitfalls if place-based data science education is to be broadly adopted as a research and classroom practice.

Authors

• Leah Rosenbaum, Teachers College, Columbia University

• Kathryn Lanouette, William & Mary

• Cody C. Pritchard, University of Tennessee

• Bradley Russell Holtz, University of Tennessee