Skip to main content

National Science Foundation Postdoctoral Fellowship

An interdisciplinary cohort of Postdoctoral scholars in data science education research advancing discovery with AI-powered tools and the All-Campus Data Science and AI Project-based Teaching and learning (ADAPT) model.

Learn about ADAPT

Project Description

With data science growing as a career path, a professional skill, and component of literacy, the NC State Data Science and AI Academy (DSA) is creating a model for training postdoctoral fellows. This program brings together mentors with varied backgrounds, professional expertise, and research interests, who may have limited experience in data science or education research. DSA offers a unique environment for exploring data science education. DSA aims to engage learners at any level and with varying career goals by offering data science courses, analytics consulting, and research enablement. DSA courses are based on the All-Campus Data Science and AI Project-based Teaching and learning (ADAPT) model, which is designed to support learners with low barriers to entry, multiple elements of choice, and high probability of success. DSA will train a cohort of postdoctoral fellows, connecting them with mentors who specialize in education research and can support them in developing their research.

DSA will create a model for training postdoctoral scholars in data science education research and successfully induct them into the data science education and STEM-Ed research community. As postdoctoral fellows enter the workforce, this project can be expected to inform the induction of data science education research, curriculum development, and student experience. The ADAPT course model, which will be continuously improved by research, can be adopted and studied at other institutions and across disciplines.

Project Details

Project Sponsor: National Science Foundation

NSF Award Number: 2222148 

Project Award: $1,249,352

Project Timeline: October 2022 – September 2026

Project Team:  

  • PI, Dr. Rachel Levy, Executive Director, Data Science and AI Academy; Professor of Mathematics 
  • Postdoctoral Fellow (2023-2025), Dr. Zarifa Zakaria
  • Postdoctoral Fellow (2024-2026), Dr. Kelsey Dufresne
  • Postdoctoral Fellow (2024-2026), Dr. Tom Leppard
  • Postdoctoral Fellow (2024-2026), Dr. Jeanne McClure 

Project Mentors: 

  • Dr. Tania Allen, Associate Professor of Media Arts, Design and Technology
  • Dr. Sunghwan Byun, Assistant Professor of Mathematics Education 
  • Dr. Ela Castellanos-Reyes, Assistant Professor of Learning of Design and Technology
  • Rob Chew, Fellow, Senior Research Data Scientist, RTI
  • Dr. Emily Griffith, Professor of the Practice, Department of Statistics
  • Dr. Yvonne Kao, Director of Learning and Technology, WestEd
  • Dr. Steve McDonald, Professor of Sociology and Anthropology
  • Dr. Gemma Mojica, Friday Institute Research Scholar
  • Dr. Jason Swarts, Professor of English 

Postdoctoral Fellows

Tom R. Leppard

Postdoctoral Research Fellow

Jeanne McClure

Postdoctoral Research Fellow

Zarifa Zakaria

Postdoctoral Research Fellow

Get To Know The Postdocs

Research Description

My research examines the capacities of experiential and experimental learning, using creative, multimodal, and project-based approaches to activate critical pedagogy. Through an interdisciplinary lens, I explore how historically situated ideologies are reflected, reinforced, or challenged in knowledge-making and sharing practices. My work engages data humanist methodologies and design frameworks, drawing on scholars such as Sara Ahmed, bell hooks, and Catherine D’Ignazio. I incorporate digital humanities, storytelling, and critical making to investigate how design, media, and technology shape lived experiences. My projects, from virtual reality reconstructions to interactive quilts and mural engagements, emphasize making as both a research method and a tool for reimagining more accessible and impactful learning environments. Through these efforts, I aim to bridge the humanities and data science, fostering critical and creative approaches to data-driven inquiry.

Presentations

August 2025

  • Presentation for LILLY Conference, Asheville, North Carolina
    • This presentation, titled “Prioritizing Creation in the Classroom through Transdisciplinarity, Storytelling, and Project-based Learning,” illustrates the utility of the ADAPT model across all disciplines and why project-based learning should be implemented in all classrooms to foster transdisciplinarity and an experiential form of learning that is rooted in multiliteracy development and practice. To showcase these affordances, this presentation explores a classroom case study of the ADAPT model in practice: the Data Diary. With the Data Diary, students learn how to engage in and with generative AI to compose visual representations of their mediated lived experiences. In doing so, they engage in a co-developed memoir-making process that enables low-stakes engagement with storytelling at the heart.

June 2025

  • Presentation for UNC System Learning and Technology Symposium, Chapel Hill, North Carolina.
    • This presentation, titled “The Stories We Share: Integrating the Humanities and Data Sciences,” provides an overview of the pedagogical model utilized by NC State University’s Data Science and AI Academy (DSA): ADAPT, the All-campus Data Science and AI Project-based Teaching and learning model. Moreover, this presentation explores how the ADAPT model and its student-centeredness and prioritization of project-based learning reinforce humanistic teaching and learning into data-science classrooms. This is illustrated by exploring a DSA course at NC State: “Storytelling with Data and AI: The Story of Self.”

May 2025

  • Presentation for LILLY Conference on Evidence-Based Teaching & Learning, Austin, Texas.
    • This presentation, titled “Engaging in Experiential Learning through Storytelling with Small Data,” explores how and why intentional, critical, and careful tending to and adoption of humanistic learning practices in increasingly pervasive data science spaces can reinforce teaching and learning that is rooted in equity, student-centeredness, and transdisciplinarity. This will be illustrated by exploring a DSA course at NC State: “Storytelling with Data and AI: The Story of Self.” This presentation empowers participants to adopt, embrace, and reimagine data science practices, principles, and products into their classrooms and workspaces to encourage multiliteracy development in an ever-increasingly data-driven world.

April 2025

  • Talk for NC State Data Science and AI Week Speaker Series
    • Dufresne’s session “Data Humanists: Small, personal, imperfect data-informed practices” played off her current research and showcased her passion for data visceralization and communication. The talk focused on data science, humanities-informed work, and personal data.
  • Presentation with Alin Yalcinkaya – The Digital Humanities Collaborative of North Carolina (DHC-NC), Digital Humanities Institute
    • This presentation, titled “Making Memoirs: Engaging Students in Storytelling through Data, AI, and Digital Humanities,” showcases the curricular design and student work of our course “Storytelling with Data and AI: The Story of Self.” This course explores what is needed to generate a good story. Throughout the course and alongside artistic AI creation and play, students engage in AI and data generation through applications such as Luma Dream Machine, Midjourney, Leonardo AI and more. In doing so, students develop a semester-long data diary rooted in reflection and AI image generation to explore the capabilities and capacities of data and AI as tools for storytelling and self-expression.
  • Training for the English Department at NC State University
    • The Data Science and AI Academy at NC State designs and delivers instruction through the ADAPT model. While DSA courses are open to all learners and have included undergraduates, graduate students, faculty, staff, alumni, and community members, the DSA’s ADAPT model serves as a valuable resource and pedagogical guide for all courses across the university. Moreover, this model of teaching and learning has strong alignment with how classes have long been taught in the humanities classroom — including a focus on projects rather than tests. This training and workshop offer an overview of the ADAPT model and opportunities for implementation in the English Department classroom.
  • Workshop for the Peer Scholars Program with NC State University Libraries
    • This workshop titled, “On Publishing in Academia: from Journals to Zines and Podcasts,” explores and tackles publishing for academia, how to establish a plan and goals for publishing, and how to ultimately share your work beyond the classroom. And while traditional journal articles, chapters, and books will be discussed, so will podcasts, zines, multimodal digital projects, galleries, and archives. By the end of this session, participants will have a list of potential avenues for publishing, sharing, and disseminating their ideas, work, and research.

March 2025

  • Presentation for NeMLA, Philadelphia, Pennsylvania
    • This presentation titled, “Drawing Up Data Sciences: Adopting ADAPT in the English Classroom,” argues that while the ADAPT model has been developed and successfully implemented across its data science courses and curriculum, such as classes on R, Python, and Data Science for Social Good, the model could be and should be adopted into English classrooms. Moreover, the ADAPT model aligns with and supports student-centered classroom strategies, which all exist beyond disciplinary confines. As such, this presentation outlines the ADAPT model, draws alignment between the ADAPT model and pedagogical theory, and grounds this educational framework within the English classroom.
  • Designer for Art In Bloom 2025, North Carolina Museum of Art
    • The particular floral arrangement Dufresne created represented “The Adoration of the Magi” by Simon Claasz. van Waterlant II, but also what data science, data-informed making, and knowledge production (as connected to a place of learning like NC State) can look like and accomplish.

October 2024

  • Presentation for the Academic Data Science Alliance, Co-led with Jeanne McClure.
    • This workshop, titled “Engaging Minds with Project-Based Learning in Data Science,” focused on the ADAPT model, the pedagogical goals and frameworks within the DSA, and how project-based learning (PBL) supports students in the classroom. Short Course: This workshop focuses on teaching educators how to structure PBL courses that enhance student creativity and agency, using clear milestones and rubrics centered around the ADAPT model. It also emphasizes inclusivity by connecting projects with students’ diverse backgrounds, promoting a more impactful learning environment.

Publications and Leadership

  • Conference Co-Organizer: 2025 Data Science and AI Education Share Fair Conference, Data Science and AI Academy, Raleigh, North Carolina
    • The 2025 Data Science and AI Education Share Fair showcases student-led projects that demonstrate the impact of the ADAPT framework in data science education. Designed for instructors, researchers, and practitioners, the event highlights how project-based learning fosters real-world skills and deep engagement with data. Attendees will explore innovative student work, exchange ideas, and gain actionable insights for integrating effective, workforce-focused teaching.
  • Care Fully: Critical Care through Making with Food (book published with NC State University Libraries and distributed by UNC Press)
    • What role do critical making, critical pedagogy and experiential inquiry play in supporting equity and justice? In considering the role of care in reimaging power, who is responsible for the care, labor, and power of ancestral and community stories, histories, and affect? Moreover, how does gender intersect with care? As seen with the capacities of cookbooks, how can systems and manners of knowing more equitably and critically enable exploration, learning, and care to reimagine power (of care, literacy, and justice) and knowledge (production, recognition, and legitimacy)? Born out of an anti-dissertation, Care Fully performs as a cookbook, working in collaboration with food makers from Raleigh, Durham, and Chapel Hill, to explore the gendered labor of making with food, the burden and responsibility of care labor, and the capitalization and fetishized commodification of care.
  • “If I Were” – an AI memoir-making project with Alin Yalcinkaya (Ongoing Project)
    • Spanning over a year, Dufresne and Yalcinkaya will create one AI-generated image to get to know each other better. Each week, they follow a unique prompt always rooted in “If I were…” Thus far, they have speculated themselves as buildings, desserts, jackets, new ages, magical creatures, their names, and more. Through this effort, they are exploring what it means to use AI as an interrelational tool, one that can help tell a story while simultaneously taking part in it.

Tom R. Leppard

Postdoctoral Research Fellow

Research Description

I study the interplay between student data science identity, classmate interactions, and pedagogical practice. My research focuses on how individuals feel a part of data science through interactions with fellow classmates in higher education data science. I am devoted to uncovering pedagogical practices that are associated with NC State’s DSA ADAPT teaching model that help facilitate the construction of professional identity and classmate interactions. Presently, I theorise that certain practices associated with project-based learning models like the ADAPT can help foster a learning framework that can ensure a positive association between classmate interactions and individual data science identity.

To learn more about Leppard’s research, visit his website.

Tom R. Leppard CV

Presentations

April 2025

  • NC State WolfWebs Talk
    • Leppard presented a talk titled “Givers, takers, and Reciprocators: Reimagining Individuals and Groups in UK Grime Music” at WolfWebs (a DSA-sponsored group of faculty and students). The talk explores using social network tools to measure an individual’s integration into a group.
  • Presentation for Symposium on Data Science and Statistics, Salt Lake City, Utah
    • Leppard presented a short talk and poster at the Statistics and Data Science Symposium based on his research into pedagogy, peer networks, and professional identity.

March 2025

October 2024

  • Presentation for 2024 Academic Data Science Alliance Annual Meeting
    • Leppard gave a lightning talk presenting his ongoing data collection efforts at the DSA. He presented his conceptual/theoretical model in a 5-minute presentation titled, “Pedagogy, Peer Networks, and Professional Identity.”

July 2024

  • DSA Postdoctoral Workshop
    • Leppard presented a research proposal to the DSA postdoctoral cohort.

Publications and Leadership

  • Conference Co-Organizer: 2025 Data Science and AI Education Share Fair Conference, Data Science and AI Academy, Raleigh, North Carolina
    • The 2025 Data Science and AI Education Share Fair showcases student-led projects that demonstrate the impact of the ADAPT framework in data science education. Designed for instructors, researchers, and practitioners, the event highlights how project-based learning fosters real-world skills and deep engagement with data. Attendees will explore innovative student work, exchange ideas, and gain actionable insights for integrating effective, workforce-focused teaching.

Jeanne McClure

Postdoctoral Research Fellow

Research Description

Dr. Jeanne McClure is a GenAI data scientist and postdoctoral fellow at NC State University’s Data Science and AI Academy. Her research explores how large language models (LLMs), retrieval-augmented generation (RAG), knowledge graphs, and AI agents can support data science instruction in both K–12 and higher education. She investigates how prompt engineering and OpenAI technologies shape student engagement, learning outcomes, and instructional design. Her work spans diverse domains, including journalism, STEM, climate science, and public health, focusing on how learners apply AI to real-world challenges. In addition to her research, she develops teaching tools and workflows and provides training and consultation on the ethical and effective use of AI. Her approach emphasizes responsible data governance and equitable access to AI technologies in education.

Presentations

April 2025

  • Short Course for Symposium on Data Science & Statistics (SDSS), Salt Lake City, Utah
    • This short course titled “Integrating Large Language Models in Introductory Data Science Courses” supports instructors in integrating large language models (LLMs) into introductory data science courses for novice programmers across disciplines. Participants explore core prompting techniques (e.g., Zero/Few Shot, Chain of Thought, Persona Prompting), compare model and interface options, and discuss effective strategies for helping students evaluate LLM-generated outputs. The course also offers guidance on ethical use, common student challenges, and includes resources to create student-facing materials and activities.

March 2025

  • Hands-on presentation for NC State Data Science and AI Week
    • This session, titled “Effective Prompting: Crafting Better Instructions for LLMs,” helps students craft effective prompts to generate personalized career development materials using large language models (LLMs). Participants will learn to create prompts that produce tailored cover letters, resumes, and slide presentations, while also using LLMs as efficient “checkers” or “raters” to improve tone, clarity, and alignment with job descriptions. The session emphasizes practical strategies students can use to build confidence and polish their professional materials quickly and effectively.
  • Hands-on presentation for AI Forward: Leading Innovation, Research and Learning at NC State University, Raleigh, NC
    • This session, titled “AI-Driven Efficiency: Mastering the Art of Strategic Prompting,” equips faculty, staff, and administrators with strategic prompting techniques to enhance efficiency and decision-making within their specific domains. Participants learn how to use large language models (LLMs) to streamline tasks like drafting, data analysis, and planning, without compromising scholarly judgment. The workshop emphasizes responsible AI use, hands-on practice, and tailored guidance for integrating prompting into academic and administrative workflows.
  • Hands-on presentation for Women in Data Science (WiDS) Conference, Raleigh, North Carolina
    • This session, titled “Harnessing Large Language Models for Data Science,” is designed for practitioners and data scientists seeking to leverage large language models (LLMs) for efficient code development, task automation, and tool integration. Participants explore advanced prompting strategies, agent workflows, and best practices for using hosted models, open-source platforms, and enterprise-grade LLMs. The workshop provides hands-on experience with techniques such as chain-of-thought reasoning and role-based prompting to accelerate development while maintaining control over data governance and model output.

February 2025

  • Presentation for AI with the FI: From Insight to Innovation in Higher Education, William & Ida Friday Institute for Educational Innovation, Raleigh, North Carolina
    • This session, titled “AI-Driven Data Visualization with Large Language Models,” introduces prompt engineering techniques, such as one-shot, few-shot, and chain-of-thought prompting, to guide large language models (LLMs) in generating and refining data visualizations. Participants learn how to structure prompts that translate raw data into clear, interpretable charts using tools like Python, R, or visualization libraries embedded within LLM workflows. The session emphasizes practical strategies for controlling output format, selecting appropriate chart types, and iterating through visual storytelling with precision and clarity.

October 2024

  • Poster Presentation for 2024 Academic Data Science Alliance (ADSA) Annual Meeting, Ann Arbor, Michigan
    • This poster titled, “From Doubt to Doing: Evolving the Data Science Framework,” presented evidence that the ADAPT model fosters our updated Data Science Literacy framework – data science literacy integrates components like data literacy, AI literacy, and others, preparing students to tackle complex, data-rich challenges across various fields. This approach, validated through interviews with students, emphasizes the role of instructors in developing these competencies, offering a comprehensive strategy for higher education institutions to equip students with essential data science skills for the future.
  • Short Course for 2024 Academic Data Science Alliance (ADSA) Annual Meeting, Ann Arbor, Michigan
    • This workshop, “Engaging Minds with Project-Based Learning in Data Science,” focuses on teaching educators how to structure PBL courses that enhance student creativity and agency, using clear milestones and rubrics centered around the ADAPT model. It also emphasizes inclusivity by connecting projects with students’ diverse backgrounds, promoting a more impactful learning environment.

September 2024

  • Workshop for Data and AI at Work, Data Science and AI Academy, Raleigh, North Carolina
    • This workshop series, titled “Building Insights from Survey and Census Data,” builds participants’ technical fluency in analyzing survey, census, and IPEDS data using R in Posit Cloud. With an emphasis on understanding community impact, the sessions guide participants through statistical techniques such as t-tests, ANOVA, chi-square tests, regression, and factor analysis. Participants explore real-world datasets to uncover patterns, address bias and missing data, and translate findings into actionable strategies for research, policy, and advocacy.

August 2024

  • Workshop for IUSE ADAPT Cohort 2 Workshop, Data Science and AI Academy, North Carolina State University
    • A professional development series was developed and facilitated to support instructors in implementing the ADAPT model through project-based learning activities at their home institutions. This multi-phase process included an orientation workshop, bi-weekly meetings, and reflective debriefs. The workshop, held in Aug 2024, helped instructors co-design inquiry-based projects and prepare for classroom implementation in Fall 2024.
  • Tutorial for useR! Conference, Durham, North Carolina
    • This tutorial, titled “R You Out of Memory Again? Level Up Your Data Game with Arrow and DuckDB,” helps tidyverse users scale their data workflows by introducing Apache Arrow and DuckDB, two powerful tools that break memory barriers while keeping your favorite dplyr syntax intact. Learn to analyze larger-than-memory datasets efficiently on your laptop using real-world examples, practical tips, and code you can immediately apply.

July 2024

  • Poster Presentation for 17th Conference of Educational Data Mining (EDM), Atlanta, Georgia
    • This poster, titled “It’s All About the Prompt: Deductive Coding’s Role in AI vs. Human Performance,” discusses how Large Language Models (LLMs) compare to human coders in deductive qualitative coding, particularly for classifying cognitive engagement in educational research. Using a mixed-methods approach, we found that LLMs, when given combination prompts, outperformed human coders by over 13% in overall accuracy and showed significant gains in identifying nuanced behaviors like Active and Constructive engagement. These findings highlight the potential of LLMs to streamline coding workflows, reduce costs, and improve consistency in complex classification tasks.
  • Presentation for the Learning Analytics in STEM Education Research Institute, Raleigh, North Carolina
    • This hands-on workshop titled “Adapt Model and PjBL Assessment” introduces participants to the ADAPT framework, emphasizing project-based learning (PjBL) to build practical data science skills. Participants move through structured milestones, from selecting a dataset to performing analysis and presenting findings, while receiving scaffolded instruction in both R and Python. The ADAPT model supports accessible, applied learning experiences that connect data science to real-world problems across disciplines.

Publications and Leadership

  • Conference Co-Organizer: 2025 Data Science and AI Education Share Fair Conference, Data Science and AI Academy, Raleigh, North Carolina
    • The 2025 Data Science and AI Education Share Fair showcases student-led projects that demonstrate the impact of the ADAPT framework in data science education. Designed for instructors, researchers, and practitioners, the event highlights how project-based learning fosters real-world skills and deep engagement with data. Attendees will explore innovative student work, exchange ideas, and gain actionable insights for integrating effective, workforce-focused teaching strategies.
  • IRB Approved Study: “Evaluating the Impact of LLM Prompt-Based Instructional Tools on Data Science Learning”
    • McClure, J., Faith, J., Ferrell, M., Zakaria, Z. & Byun, S. (April 2025). Evaluating the Impact of LLM Prompt-Based Instructional Tools on Data Science Learning assesses the effectiveness of structured “Recipe Books” (tailored prompt-based instructional tools) in enhancing student learning when using Generative AI tools for diverse data science courses.
  • LASER Principles for Peer Review
    • McClure worked on updating the LASER principles to develop a peer review framework for curriculum and instructional materials to better align with ADAPT, UDL 3.0, Universal Design, and Design Justice Principles. This revision integrates hands-on, project-based learning with support for inclusive technical expression and multiple means of engagement. It emphasizes flexibility in how students demonstrate knowledge, including coding alternatives and assistive technologies. The updated principles ensure materials are both workforce-relevant and accessible to diverse learners.
  • Developer and Instructor: Learning Analytics Workflow Course. Learning Analytics in STEM Education Research Institute, Raleigh, North Carolina
    • McClure, J. (2024). Developed Learning Analytics Workflow modules in both R and Python to support STEM education researchers in analyzing educational data to enhance student learning. These modules combine foundational theory with hands-on coding exercises, guiding participants through data preparation, visualization, modeling, and communication. Drawing from key texts and current research, the modules provide practical experience tailored to varying programming backgrounds.
  • Journal Article: “Toward inclusivity in Artificial Intelligence: A comparative study of cognitive engagement between marginalized female students and peers. British Journal of Educational Technology
    • Jiang, S., McClure, J *, Tatar, C. *, Bickel, F. *, Rosé, C., & Chao, J. (2024). This study explores how high school students engage cognitively with machine learning practices, focusing on differences in learning experiences and outcomes. Researchers analyzed responses from 27 students using statistical tests, content analysis, and interviews to examine how students developed machine learning models. Findings showed that students who incorporated a range of cultural perspectives demonstrated stronger performance, particularly in tasks involving holistic language analysis and interpreting writer intent. The study highlights the value of integrating language analysis and machine learning across subjects to support student engagement and learning.
  • Study: Dynamics of Distance Education institutions with high enrollment rates. National Center for Educational Statistics Annual Meeting (NCES), Washington, DC
    • Armstrong, E., Grafe, C., Liang, Y., McClure, J., Tegen., M.* (June, 2024). This exploratory study examined graduation outcomes at high-enrollment Distance Education (DE) institutions using NCES data. Findings suggest shifting demographic trends, with younger students contributing to improved graduation rates in recent years. While total undergraduate enrollment showed mixed associations with success, some institutions stood out, possibly due to innovative models like Competency-Based Education, highlighting the need for further research into targeted supports grounded in Adult Learning Theory.
  • Book Chapter: “Embracing Learning Analytics in Health Professions Education,” New Directions for Teaching and Learning
    • Hibbard, S., McClure, J., & Kellogg, S. (2024). This chapter explores how learning analytics can be leveraged to improve teaching and learning in health professions education. It highlights practical applications of data-driven methods, such as machine learning, network analysis, and simulation tracking, to support curriculum design, clinical training, and personalized learning. Readers will also gain insight into key methodologies, data sources, and ethical considerations shaping the future of analytics in healthcare education.
  • Conference Paper: “Navigating Cognitive Engagement in AI-Enhanced Education: Lexical Diversity and Open-Ended Inquiry in Journalism Learning.” Proceedings of the 17th International Conference of the Learning Sciences (ICLS), Buffalo, New York
    • McClure, J., J., Bickel, F., Jiang, S., Chao, J., Rosé, C. (2024). This study explores the relationship between lexical diversity and cognitive engagement in an AI-enhanced journalism course for high school students. Using a mixed methods approach, researchers identified distinct student clusters that revealed varying patterns of engagement linked to lexical diversity. The findings highlight that while higher lexical diversity often aligns with active engagement, it is not a sole predictor, underscoring the need for instructional strategies that foster both rich language use and deeper cognitive involvement.
  • Conference Paper: “Modeling with Primary Sources: An Approach to Teach Data Bias for Artificial Intelligence and Machine Learning Education” Proceedings of the 17th International Conference of the Learning Sciences (ICLS), Buffalo, New York
    • McClure, J., Zheng, J., Bickel, F., Jiang, S., Rosé, C., Chao, J. (June, 2024). (Awarded Outstanding Design Paper). The study introduces the SourceML approach for teaching data bias through primary sources in high school history classrooms. It uses a web-based tool, StoryQ, allowing students to build ML models with historical texts, enhancing their understanding of data bias origins, societal impacts, and mitigation strategies. This method emphasizes historical thinking, helping students recognize how biases are formed by the context of data generation and influence societal perceptions, thereby preparing them for ethical interactions with AI technologies.

Zarifa Zakaria

Postdoctoral Research Fellow

Research Description

Dr. Zakaria’s research focuses on enhancing teaching and learning experiences in STEM+C educational technologies. Her Ph.D. work concentrated on improving collaboration quality in elementary computer science classrooms and exploring the motivation of underrepresented students in STEM. Dr. Zakaria has consistently utilized mixed methods research with multimodal data to investigate educational technologies and their impact on learning. As a postdoctoral researcher at the NC State Data Science Academy, she continues to apply this approach to explore the teaching-learning dynamics in data science courses, emphasizing the integration of project-based learning and thoughtful dataset practices to improve student engagement and outcomes. Her interests also include the integration of large language models (LLMs) in higher education and the shifts in teaching-learning dynamics they create for instructors and researchers to consider.

Presentations

April 2025

  • Presentation for Data Science Education Speaker Series
    • This presentation, titled “Teaching with Data: Navigating Dataset Choices in Higher Education,” explores dataset curation as a foundational yet often overlooked instructional practice in introductory data science, drawing on focus group interviews with eight instructors to examine how they select datasets that balance accessibility, learning affordances, and student relevance. We share key findings, highlight tensions instructors face in dataset selection, and invite participants to collaboratively explore how to better support and study this critical aspect of data science pedagogy.

July 2025

  • Presentation for Computer Science Teachers Association (CSTA) 2025 Annual Conference
    • This presentation, a collaboration with Dr. James Harr titled, “From Numbers to Stories: Creating Visualizations with Introductory Python & R” equips middle and high school educators to teach data visualization through Python and R, using real-world datasets relevant to teens—like social media trends, climate patterns, and sports stats—to help students turn raw data into compelling visual stories. Grounded in research-backed strategies like pair programming and scaffolded challenges, the session provides ready-to-use code templates, a portfolio-ready project, curated datasets, and CSTA-aligned curriculum guides to make programming both accessible and creatively engaging.

October 2024

  • Presentation for 2024 Academic Data Science Alliance (ADSA) Annual Meeting, Ann Arbor, Michigan
    • This presentation, titled, “Enhancing Learning through Data Science Competitions: Insights from a Hackathon” focuses on data science competitions like hackathons provide immersive, hands-on learning experiences that enhance students’ skills and knowledge, and this presentation examines how one such event supported collaborative learning, skill development, and the application of prior knowledge in new contexts. Through video documentation, participant surveys, and analysis of the competition’s instructional structure, this study reveals how minimal but well-designed guidance can foster effective learning, offering insights into optimizing short-term data science competitions for educational impact.
  • Presentation for 2024 Academic Data Science Alliance (ADSA) Annual Meeting, Ann Arbor, Michigan
    • This presentation, a collaboration with Dr. James Harr and Jajwalya Karajgikar titled “Critical Making as Pedagogical Approach to the Data Science Classroom,” introduces data physicalization as both a form of critical making and a pedagogical tool that represents data through tangible, physical forms to support multimodal learning. Participants will explore its historical context, pedagogical potential in data science education, and engage in a hands-on workshop to create scalable physicalizations for their own classrooms.

May 2023

  • Presentation for 2023 Symposium on Data Science and Statistics, St. Louis, Missouri
    • This presentation, titled “What are the Characteristics of Effective Datasets for Data Science Courses?” focused on the characteristics of effective datasets in data science courses, drawing on thematic analysis of instructor interviews to highlight key elements such as balanced complexity, student relevance, and real-world connection. I share insights from focus group discussions with data science instructors on dataset selection strategies, challenges, and pedagogical approaches, offering practical guidance for educators designing data-driven learning experiences.

October 2023

  • Presentation for 2023 Academic Data Science Alliance (ADSA) Annual Meeting, San Antonio, Texas
    • This presentation, titled “Designing Project-Based Data Science Learning to Enhance Health and Well-being,” Zakaria collaboratively helped develop and present the ADAPT model, highlighting how project-based data science education can support student identity, equity, and wellness. She contributed to framing discussions and activities that encouraged educators to reflect on inclusive practices and reduce stress for diverse learners.

Publications and Leadership

  • Conference Co-Organizer: 2025 Data Science and AI Education Share Fair Conference, Data Science and AI Academy, Raleigh, North Carolina
    • The 2025 Data Science and AI Education Share Fair showcases student-led projects that demonstrate the impact of the ADAPT framework in data science education. Designed for instructors, researchers, and practitioners, the event highlights how project-based learning fosters real-world skills and deep engagement with data. Attendees will explore innovative student work, exchange ideas, and gain actionable insights for integrating effective, workforce-focused teaching strategies.
  • Conference Co-Organizer: 2024 Data Science and AI Education Share Fair Conference, Data Science and AI Academy, Raleigh, North Carolina

Postdoctoral Duties and Responsibilities

Essential Job Duties

Scholars will develop a research project focused on teaching and learning in progress through the Data Science and AI Academy at NC State University. The research will investigate some aspects of the ADAPT model (described above and online). 

The scholars will participate as a cohort in research and professional development activities such as weekly mentor meetings, leadership development, teaching meetings, a data science education journal club, brown bag lunches, presentations, workshops, and a data science education conference. Fellows may also access NC State professional development programs and funding for travel and professional memberships.

Each postdoctoral scholar will be expected to teach one course per semester for the Data Science Academy at the undergraduate or graduate level. They may develop a new course or teach an existing course.

The postdoctoral scholars will provide input that helps identify a primary and secondary research mentor.  The goal is to create interdisciplinary teams.

NC State University and the Data Science and AI Academy are especially interested in qualified candidates who can make contributions through their experience, research, teaching, and/or service that demonstrate NC State’s goals and values as described in the Wolfpack 2030 Powering the Extraordinary Strategic Plan.

Other duties as assigned.

Minimum Experience/ Education

Ph.D. or equivalent doctorate (e.g., M.D., D.V.M., Sc.D.) in an appropriate field awarded no more than two (2) years from the initial date of postdoctoral appointment.

Required Skills

  • Must be a U.S. citizen, national, or permanent resident when the application is submitted. 
  • Must have earned a doctoral degree, or expect to have earned the doctoral degree in a field of Science, Technology, Engineering, or Mathematics (STEM), STEM Education, Education, or a related discipline prior to the start date of the position. The start date must be no later than two years after the conferral of the doctoral degree.
  • Must not hold a tenure-track position.
  • Experience with data science, data science education and/or data science education research.
  • At least 3 years of relevant teaching experience in high school, postsecondary level or as an instructional trainer.
  • Relevant research experience.  We encourage applications from a broad range of disciplines, including statistics, mathematics, computer science, engineering, policy, ethics, communications, arts and humanities, physical and social sciences, or related fields.
  • Strong communication and writing skills

Qualified applicants may not have all of the following

  • Record of research or professional publication and presentation, commensurate with experience.
  • Prior work in research projects with human subjects.
  • Experience with accessibility practices in communication.
  • Experience with project-based teaching and learning.
  • Experience broadening participation in STEM education and STEM fields.
  • Experience with quantitative and/or qualitative data analysis and communication tools.
  • Demonstrated success working in collaborative teams.

Acknowledgement: This material is based upon work supported by the National Science Foundation under Grant No. 2222148 STEM Ed PRF: Mentoring a Diverse Cohort of Postdoctoral Scholars in Data Science Education Research. 

Disclaimer: Any opinions, findings, and conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.