Welcome to the official GitHub organization for The Center for Educational Data Science and Innvoation! This is the central hub for all projects, tools, and initiatives driven by The Center’s team, contributors, and community. Whether you are a developer, researcher, partner, or supporter, here you will find valuable resources, codebases, and collaborative projects dedicated to advancing our mission.

The Center for Educational Data Science and Innovation (EDSI) at the University of Maryland (UMD) is a pioneering interdisciplinary research center dedicated to transforming education through cutting-edge data science, artificial intelligence (AI), and innovative research. Housed within UMD’s College of Education, EDSI is committed to advancing education equity, effectiveness, and efficiency by leveraging the power of data and AI technologies.

Our mission focuses on three core objectives:

• Accelerating innovative research that addresses critical educational challenges by integrating data science and AI methods to promote equity and improve learning outcomes for all students.

• Bridging educators, researchers, policymakers, and industry leaders to foster the ethical, evidence-based, and responsible use of data and AI in education.

• Transforming education systems by developing technological advancements and policy solutions that enhance educational equity, efficiency, and effectiveness.

EDSI leads groundbreaking projects such as collecting high-quality multimodal classroom datasets—including anonymized transcripts, audio, video, student demographics, achievement data, and surveys—combined with human annotations of teaching practices and student mathematical engagement. This rich, user-friendly data infrastructure supports AI model development and innovative educational tools aimed at high-impact, equity-focused teaching and learning improvements.

Committed to collaboration, EDSI works closely with diverse partners from across the University of Maryland, local school districts, education agencies, policymakers, and industry sponsors—including Google and the Chan Zuckerberg Initiative—to implement data-driven strategies that improve educational experiences and outcomes. Through research, community engagement, and interdisciplinary partnerships, we strive to create equitable, effective, and efficient learning environments that benefit all students, particularly those from historically marginalized groups.

Our vision is to ethically harness data science and AI to shape the future of education, ensuring opportunities for all learners to thrive in an AI-enhanced world. Our GitHub organization serves as a transparent and collaborative space to develop, share, and maintain software and data projects that align with these goals.

EDSI is proud to highlight the release of pairadigm, a new open-source Python library developed by our Data Scientist, Michael Leon Chrzan.

pairadigm introduces a concept-guided chain-of-thought approach to pairwise comparison, enabling large language models (LLMs) to produce more reliable and scalable measurements of abstract constructs such as reasoning, creativity, toxicity, and emotion.

Key contributions of the package include:

• A unified interface for multiple LLM backends (GPT, Claude, Gemini, HuggingFace, and more).
• Psychometric scoring via Bradley–Terry models to generate continuous scales.
• Built-in validation tools, including transitivity checks and inter-rater reliability metrics, to align LLM outputs with human annotations.
• Reward modeling functionality that fine-tunes encoder models for efficient large-scale measurement.

This work bridges computational social science, education research, and NLP, offering researchers a streamlined workflow for annotation, validation, and model training.

📄 See more here: https://github.com/mlchrzan/pairadigm

• Core Projects: Maintained repositories that form the backbone of The Center’s technical efforts.
• Community Contributions: Open-source projects and tools built or supported by our contributors.
• Documentation & Resources: Guides, API documentation, datasets, and other resources to help you understand and engage with our initiatives.
• Research & Innovation: Cutting-edge experimental code, prototypes, and research implementations.
• Collaboration & Support: Issue trackers, discussions, and project boards to facilitate community involvement and support.

We warmly welcome contributions from the community. Please adhere to the following principles:

• Code of Conduct: Respectful and inclusive behavior is mandatory. Please see CODE_OF_CONDUCT.md in each repo.
• Pull Requests: Follow the Pull Request template, write clear commit messages, and provide relevant tests/documentation.
• Issue Reporting: Use issue templates where available. Provide clear, detailed information to help us address your concerns quickly.
• Licensing: All projects are under [specify license, e.g., MIT License]; please review before contributing.

If you have questions, need support, or want to collaborate:

• Open an Issue or Discussion in the relevant repository.
• Reach out via email: edsi@umd.edu.

We are excited to have you join our community and help build impactful projects together!

Follow The Center on social media and subscribe to our newsletters (links available on the main website) to stay updated on latest announcements, releases, and events.

• X: https://x.com/EDSI_UMD
• Bluesky: https://bsky.app/profile/edsi-umd.bsky.social
• LinkedIn: https://www.linkedin.com/company/center-for-educational-data-science-and-innovation/
• YouTube: https://www.youtube.com/@edsiumd

Thank you for visiting The Center’s GitHub organization. Together, let's innovate and make a difference!

The EDSI Team