Stephanie Shipp, Jing Cao, Sarah Kalicin, Edward Mirielli, Madi Thompson, Sungwhan Byun, and Adam Loy
The 2025 Symposium on Data Science and Statistics was held in Salt Lake City, Utah, from April 28 to May 2, bringing together data scientists, computer scientists, and statisticians to exchange ideas. The theme, “Bridging Disciplines: Advancing AI, Statistics, and Data Science Together,” ran throughout the conference in the plenary talks, presentations, and discussions.
The symposium welcomed 360 participants, including 111 students. It was filled with networking opportunities, presentations, lightning talks, and e-posters.
Short Courses
The symposium started with short courses organized by Ed Mirielli from the University of Missouri. These courses aimed to equip participants with new data science skills before the main conference. Diverse learning opportunities were offered to enhance participants’ knowledge and practical skills. The topics were based on suggestions and interests of ASA members. Overall, the goal of these short courses was to foster lifelong learning and professional growth.
The following short courses were offered:
- Time Series Applications with a Machine Learning Framework (full day) taught by Sean McMannamy, Graceland University
- Introduction to Interpretable Machine Learning Using SHAP, GINI, and LIME (half day) taught by Debarshi Datta, Florida Atlantic University
- Integrating Large Language Models in Introductory Data Science Courses (half day) taught by Jeanne McClure, Sunghwan Byun, Zarifa Zakaria, and Matthew Ferrell, North Carolina State University, and Joe Faith, Harding University
- Building Containerized Applications for Data Science (Part 1, half day) and Accelerating Data Science Workflows with Kubernetes (Part 2, half day), taught by J. Alex Hurt, University of Missouri
These short course sessions were well attended. Suggestions for next year’s courses can be sent to the SDSS 2026 chair, Jing Cao.
Plenary Speakers
The symposium featured three keynote speakers who highlighted advances, challenges, and the leadership and technical roles of data scientists, statisticians, computer scientists, and artificial intelligence experts.
Julia Silge from Posit opened the conference about advances in data science tools. She discussed the development of Positron, designed to support both R and Python for data analysis. Silge emphasized that Positron is built to be extensible and user-friendly, tailored specifically for data science tasks. She also addressed the tool’s ability to work with multiple programming languages and its integration with other tools like Arc and Air. Finally, Silge noted that Positron is not intended as a general-purpose software engineering tool, but rather as a solution designed for data science needs.
Looking ahead, the Posit team plans to roll out AI and large language model integration features in Posit Beta this summer. They are also working on a ‘bring your own model’ functionality for AI/LLM integration and exploring cloud version options for academic and teaching purposes, based on community interest. Additionally, the team is finalizing partnerships with AI model providers.
On the symposium’s second day, a keynote panel of the following three speakers from academia and a national lab discussed the challenges of AI and LLMs, focusing on their impact on students’ statistical education and research:
- Judith Canner, California State University-Monterey Bay
- Jimmy Doi, Cal Poly
- Karl Pazdernik, Pacific Northwest National Laboratory
The panel emphasized that researchers using AI must consider key ethical issues, including fairness, transparency, and accountability. To prepare future AI data scientists for these challenges, ethical reasoning should be integrated into technical education through real-world case studies, interdisciplinary collaboration, and hands-on tools for evaluating model effects. This approach fosters critical thinking and responsible innovation, equipping students to navigate the complex ethical landscape of AI.
Members of the panel suggested that to still be competitive in the rapidly evolving field of AI and statistics, students should build a solid foundation in mathematics, programming, and data analysis, while also gaining expertise in machine learning and statistical modeling. Traditional statistical courses—such as experimental design and sampling—can be especially valuable when taught in the context of validating AI models and ensuring their reliability. Integrating these core concepts with practical AI applications helps students develop both technical proficiency and the critical thinking needed to adapt to new challenges in the field.
On the third day, keynote speaker and 14th president of Brigham Young University Shane Reese emphasized the critical role statisticians and data scientists play in leading organizations. He highlighted how their skills can address key issues in higher education and beyond. Drawing from his background in statistics, Reese introduced the concept of “leadership through a stochastic lens,” showing how statistical thinking can enhance leadership practices. He gave many examples of how BYU uses data science and statistical analysis to address university challenges.
These examples include Reese’s approach to these challenges through a stochastic lens. He highlighted the importance of mental health support, showcasing a new freshman course to foster connections and improved counseling services that reduced wait times from three weeks to one. Additionally, they implemented a data-driven strategy to identify high-risk students for timely intervention. Reese also discussed attracting Kevin Young as the new basketball coach with the promise of analytics. His choice not only resulted in a successful season but also created a data pipeline for real-time coaching insights.
In summary, Reese emphasized the crucial role of mentorship, accountability, and humility in fostering leadership and trust. Following are the tenets he believes leaders should practice:
- A leader envisions other’s potential, rather than just their current state, which can significantly alter how they perceive themselves.
- Extreme accountability emphasizes the importance of holding oneself and others accountable to drive results. This theory emphasizes the importance of striking a balance between accountability and empathy, suggesting that acknowledging others’ strengths is crucial for effective leadership.
- Humility and meekness, distinct from weakness, can lead to greater success and satisfaction.
Special Session
A special panel session, Cultivating a Data Culture for Business Impact, was organized by Sarah Kalicin to discuss the importance of data culture in data science and AI and how statisticians and data scientists can help lead this transformation. The panelists were Lindsey Zuloaga, a vice president at HireVue; Jordan Morrow, known as the “Godfather of Data” literacy; Matt Housely, co-author of Fundamentals of Data Engineering; and Sarah Kalicin, founder and data strategist at Achieve More with Data. David Corliss of Grafham Analytics chaired the panel.
The panel discussed the definition of a culture and why a strong data culture is important to our careers. Data and business organizations use different languages and processes to do our work, which need to be brought together for a common purpose. A data culture refers to the way these groups communicate with each other to create shared organizational values, beliefs, and norms as they use data to solve business problems. The panel further discussed and provided examples of how statisticians and data scientists can help bridge these gaps by understanding the business, collaborating with key stakeholders to use data to solve their problems, and ensuring they understand the impact of their work on the business. Networking across the organization and providing self-serve data tools are key to build relationships and trust with organizational partners. These can establish a strong foundation to create the necessary common language, belief system, and norms around the use of data.
Engaging Students
SDSS included 111 students actively taking part in a range of activities, including lightning sessions, e-posters, mentoring opportunities, and social gatherings. This participation can be traced back to Madi Thompson (SDSS committee member), who reached out to universities across Utah. Many educators responded, securing funds to enable their students to showcase their work and gain experience. Adam Loy, Sungwhan Byun, and Ed Mirielli (SDSS committee members) helped create an organized meetup. Students gathered in small groups to share insights from the conference, fostering new connections and collaborative ideas. The evening continued with Brian Knaeble from Utah Valley University leading a hike near the University of Utah.
View more SDSS photos on Flickr.
Save the Date for SDSS 2026 in Milwaukee!
SDSS 2026 is set to take place in Milwaukee, Wisconsin, from April 27 to May 1. Under the leadership of Jing Cao, the chair, and Sarah Kalicin, chair-elect, plans are underway. Milwaukee is situated on the shores of Lake Michigan, offering access to parks, beaches, and a range of outdoor activities. The city has many attractions, including museums, theaters, sports venues, and a culinary scene. Mark your calendars for SDSS 2026 and be sure to watch for abstract and other deadlines in Amstat News.
Leave a Reply