Loading...
Organizing Committee
Abstract
As the field of data science has expanded, there has been a notable increase in collaboration between computer science, statistics, and applied mathematics. This interdisciplinary approach has influenced the structure of traditional statistics education. Specifically, there is a growing emphasis on incorporating computer science skills, such as programming, machine learning, and computational tools for managing big data, into statistics curricula. However, the role of applied mathematics in statistics and data science education often remains under-discussed. This workshop aims to highlight the importance of mathematical modeling, optimization, simulation, and prescriptive analytics in complementing a statistical or data science education.
This workshop is designed for faculty, postdocs, and PhD students who are interested in teaching statistics and data science through an interdisciplinary approach that leverages the strengths of both applied mathematics and statistics. Participants will exchange pedagogical ideas, curricula, and research applications, fostering collaboration between those teaching data science. The workshop is especially relevant for statistics faculty looking to expand their data science toolbox and applied mathematicians aiming to equip their students with essential data science skills. The workshop is also open to individuals in engineering, epidemiology, computer science and other STEM or public health fields.
This workshop is designed for faculty, postdocs, and PhD students who are interested in teaching statistics and data science through an interdisciplinary approach that leverages the strengths of both applied mathematics and statistics. Participants will exchange pedagogical ideas, curricula, and research applications, fostering collaboration between those teaching data science. The workshop is especially relevant for statistics faculty looking to expand their data science toolbox and applied mathematicians aiming to equip their students with essential data science skills. The workshop is also open to individuals in engineering, epidemiology, computer science and other STEM or public health fields.
Loading...
Loading participant list in background...
Workshop Schedule
Monday, May 19, 2025
8:30 - 8:50 AM EDT
Check In
11th Floor Collaborative Space
8:50 - 9:00 AM EDT
Welcome
11th Floor Lecture Hall
Brendan Hassett, ICERM/Brown University
9:00 - 9:20 AM EDT
Saving Lives on Planet Earth
11th Floor Lecture Hall
Speaker
Eva Lee, The Center for Operations Research in Medicine and Healthcare
Session Chair
Alyssa Bilinski, Brown University
Abstract
Mathematical modeling offers a versatile framework for understanding, predicting, and analyzing real-world systems by translating complex phenomena into simplified explainable mathematical representations. This allows for the testing of scenarios and designing solutions, making predictions, optimizing systems and processes, and informing decisions in various fields, from science and engineering to social sciences and public policy. In this talk, we will discuss several real-world mathematical models covering pathogens, behavior, and logistics operations, and their impacts to strategic policies, tactical actions, and operational decisions.
9:30 - 9:50 AM EDT
Latent-Variable Modeling of Unmeasured Confounding Using Proxy Variables
11th Floor Lecture Hall
Speaker
Amy Cochran, University of Wisconsin-Madison
Session Chair
Alyssa Bilinski, Brown University
Abstract
Every day, emergency department providers make critical decisions—like whether to admit a patient—based on limited information. Researchers analyzing these decisions from electronic health records (EHR) face a similar challenge: crucial factors affecting decisions and outcomes are often unmeasured. Many standard statistical methods assume all relevant confounders are measured, leading to potentially misleading results. To solve this problem, we developed a mathematical modeling approach that uses proxy variables—data readily available in the EHR, such as vital signs—to indirectly estimate these otherwise latent confounders. Across multiple studies, our method more thoroughly accounts for confounding compared to standard analyses, frequently reversing their conclusions (e.g., demonstrating that admission is beneficial rather than harmful). Motivated by these findings, we created a broadly applicable, proxy-based framework that is both practical and robust to model misspecification. Our method offers researchers a novel and accessible tool for addressing unmeasured confounding in healthcare settings beyond emergency medicine.
10:00 - 10:30 AM EDT
Coffee Break
11th Floor Collaborative Space
10:30 - 10:50 AM EDT
Napkin Math: Why Simple Models Matter (More) in an Era of High-Performance Computing
11th Floor Lecture Hall
Speaker
Alyssa Bilinski, Brown University
Session Chair
Alice Paul, Brown University
Abstract
Over the past few decades, researchers have seen rapid advancements in computational power, allowing for the rapid development and dissemination of complex statistical and mechanistic simulation models. Amidst this backdrop, I will argue that simple models remain critical – both on their own and as complements to more complicated “black boxes.” I will begin by highlighting features that differentiate policy modeling from other prediction problems, including discrete (often binary) decisions; diverse, user-specific objectives; and asymmetric costs between false positives and false negatives. I will then illustrate, with examples from disease simulation modeling, techniques for employing simple models, such as "aiming off" and benchmarking.
11:00 - 11:20 AM EDT
The Role of Simulation in Data Science
11th Floor Lecture Hall
Speaker
Reza Yaesoubi, Yale
Session Chair
Alice Paul, Brown University
Abstract
In this talk, we will review the applications of simulation modeling in data science. There are many situations where the data to study the questions of interest are expensive, infeasible, or unethical to collect. For these cases, we can rely on simulation modeling to describe the underlying real-world system and generate the data that we need for our analysis. We will review the basic concepts of simulation modeling, a number of applications, some challenges, and future directions.
11:30 - 11:50 AM EDT
Constrained Optimization with Sample Estimates of Parameters
11th Floor Lecture Hall
Speaker
James Cochran, The University of Alabama
Session Chair
Alice Paul, Brown University
Abstract
The traditional constrained optimization paradigm considers problems to be either deterministic (the values of all parameters are known with certainty) or stochastic (the value of at least one parameter is not known with certainty, but its probability distribution is known).We expand this paradigm by considering a third class of constrained optimization problems - problems for which the value of at least one parameter is not know with certainty and the probability distribution is not known, and so the value of the parameter is estimated using sample data. We also consider Bayesian and predictive approaches to these problems and bootstrap approaches to inference on the estimated maximand or minimand for such problems. We will also provide a few interesting examples.
12:00 - 12:20 PM EDT
International Vaccine Allocation: An Optimization Framework
11th Floor Lecture Hall
Speaker
Michael Veatch, Gordon College
Session Chair
Alice Paul, Brown University
Abstract
A multi-area SIR (susceptible-infectious-recovered) epidemiology model is embedded in a optimization model with vaccination allocation decisions across areas. A more contagious variant emerges after a random number of person-infectious days. After several approximations, the resulting nonconvex program is solved using a sequence of linear program approximations. The model uses COVID-19 data to identify scenarios under which it might be in a high-income nation's own interest to donate vaccine doses to another country before its own population has been fully vaccinated, postponing the variant. Generally, donating vaccine greatly decreases total predicted deaths, with a modest or no increase in donor country deaths. The model uses global data and up to 10 areas, representing groups of countries. It could be used with more areas and country-specific data.
12:30 - 1:30 PM EDT
Networking Lunch
Working Lunch - 11th Floor Collaborative Space
1:30 - 1:50 PM EDT
Finding Data in Unexpected Places
11th Floor Lecture Hall
Speaker
Sheldon Jacobson, University of Illinois at Urbana-Champaign
Session Chair
Alyssa Bilinski, Brown University
Abstract
The growth of artificial intelligence has made data a valuable asset. Though data has become ubiquitous, sometimes there are unexpected and untapped data sources waiting to be discovered. This presentation reports an exercise with banknotes printed by the United States Bureau of Engraving and Printing. Given that every banknote contains a unique eight-digit serial number, the objective is to build models that inform banknote collectors how much effort they must expend to find banknotes with matching serial numbers. These models generalize a classic problem from discrete mathematics, termed the sock matching problem. Analytical and computer simulation models provide insights that inform how much data is needed to obtain the desired prize. This talk provides a glimpse of new data opportunities exist that may be flying “under the radar”.
2:00 - 2:20 PM EDT
Discrete Optimization Algorithms for Estimation
11th Floor Lecture Hall
Speaker
Alice Paul, Brown University
Session Chair
Alyssa Bilinski, Brown University
Abstract
Risk scores are sparse linear models that map an integer linear combination of covariates to a probability of an outcome occurring. They are a popular predictive model in healthcare because they are easy to use and interpret, allowing clinicians to calculate a patient's risk for a given outcome by hand. This talk focuses on the underlying maximum likelihood estimation of these models, which involves a discrete optimization problem. We compare different algorithmic approaches and the trade-offs between computational efficiency and performance for this problem. Last, we highlight other recent applications of optimization in best subset selection and demand estimation in dockless mobility.
2:30 - 2:50 PM EDT
Operations Research in Action
11th Floor Lecture Hall
Speaker
Anna Svirsko, United States Naval Academy
Session Chair
Alyssa Bilinski, Brown University
Abstract
The traditional Operations Research (OR) curriculum is designed to train students to use mathematical modeling to solve real-world problems. In today’s data-driven landscape, this means equipping students to integrate large datasets and OR techniques to address problems and communicate insights to decision makers. In this talk, I will present the newly piloted Operations Research in Action course at the United States Naval Academy. This course is designed to guide students through the entire problem solving process, from data exploration and model development to stakeholder communication, to ultimately transform them into effective OR analysts.
3:00 - 3:30 PM EDT
Coffee Break
11th Floor Collaborative Space
3:30 - 4:00 PM EDT
Paired Curriculum Review
Group Work - 11th Floor Lecture Hall
4:00 - 4:30 PM EDT
Light Reception
Reception - 11th Floor Collaborative Space
4:30 - 5:30 PM EDT
Journey of a Black Mathematician
Optional Screening - 11th Floor Lecture Hall
Tuesday, May 20, 2025
9:00 - 9:20 AM EDT
From Northampton to Queens and back
11th Floor Lecture Hall
Speaker
Ben Baumer, Smith College
Session Chair
Anarina Murillo, Brown University
Abstract
After completing a master's degree in applied math at the University of California, San Diego in 2003, I spent about 12 years in the academic wilderness before settling on a professional identity as a tenure-track data scientist at Smith College. Along the way, I learned a lot about how important (and unimportant) our professional identities can be, especially in academia. In this talk, I discuss some of these lessons learned.
9:30 - 9:50 AM EDT
Everything is Knowledge
11th Floor Lecture Hall
Speaker
Kun Meng, Brown University
Session Chair
Anarina Murillo, Brown University
Abstract
My background is a mixture of statistics, biostatistics, applied mathematics, pure mathematics, and data science, which provides me with the experience of being taught by professors with different expertise and the opportunity to teach students with different interests. The main theme of my teaching philosophy, shaped by my diverse background, is that I do not compartmentalize material into separate categories such as mathematics or data science—instead, all are under the same category: knowledge. I will illustrate this philosophy with a brief teaching demonstration.
10:00 - 10:20 AM EDT
Functions are all you need
11th Floor Lecture Hall
Speaker
Nathan Carter, Bentley University
Session Chair
Anarina Murillo, Brown University
Abstract
My dissertation was in mathematical logic, with no exposure to statistics or applied mathematics. As I was finishing a sabbatical in 2019, my chair told me that the person who had taught our data science courses left for industry, so I'd been assigned a graduate-level data science course that fall. Oh, and welcome back! Fortunately, mathematics is a strong preparation for many related fields, including data science. In this talk, I'll showcase how the simple concept of a function underpins huge areas of data work, from simple modeling to AI chatbots. In particular, the robust familiarity with functions that mathematics requires (e.g., complex domains and codomains, function spaces, etc.) lays the perfect foundation.
10:30 - 11:00 AM EDT
Coffee Break
11th Floor Collaborative Space
11:00 AM - 12:00 PM EDT
Teaching Programming
Panel Discussion - 11th Floor Lecture Hall
Session Chair
Alice Paul, Brown University
Panelist
Ben Baumer, Smith College
Panelist
Mine Cetinkaya-Rundel, Duke University
Panelist
Anarina Murillo, Brown University
Panelist
Alex Reinhart, Carnegie Mellon University
12:00 - 12:05 PM EDT
Group Photo (Immediately After Talk)
11th Floor Lecture Hall
12:05 - 1:00 PM EDT
Networking Lunch Provided by ICERM
Working Lunch - 11th Floor Collaborative Space
1:00 - 1:05 PM EDT
Help from AI
Lightning Talks - 11th Floor Lecture Hall
Speaker
Mine Cetinkaya-Rundel, Duke University
Session Chair
Alice Paul, Brown University
1:05 - 1:10 PM EDT
How I Teach
Lightning Talks - 11th Floor Lecture Hall
Speaker
Eva Lee, The Center for Operations Research in Medicine and Healthcare
Session Chair
Alice Paul, Brown University
1:10 - 1:15 PM EDT
Artisanal Programming
Lightning Talks - 11th Floor Lecture Hall
Speaker
Ben Baumer, Smith College
Session Chair
Alice Paul, Brown University
1:15 - 1:20 PM EDT
Leveraging data technologies + SQL to bring bigger data into the classroom
Lightning Talks - 11th Floor Lecture Hall
Speaker
Nicholas Horton, Amherst College
Session Chair
Alice Paul, Brown University
1:20 - 1:25 PM EDT
How I teach potential outcomes (algorithmically)
Lightning Talks - 11th Floor Lecture Hall
Speaker
Amy Cochran, University of Wisconsin-Madison
Session Chair
Alice Paul, Brown University
1:25 - 1:30 PM EDT
How I use disaster response to teach about values in modeling
Lightning Talks - 11th Floor Lecture Hall
Speaker
Michael Veatch, Gordon College
Session Chair
Alice Paul, Brown University
2:00 - 2:20 PM EDT
The Future of Statistics Education: A Computational Perspective
11th Floor Lecture Hall
Speaker
Mine Cetinkaya-Rundel, Duke University
Session Chair
Anarina Murillo, Brown University
Abstract
Statistics education stands at a critical juncture as we navigate the intersection of traditional statistical theory, modern computational approaches, and emerging AI technologies. This talk examines how statisticians can reimagine curricula by embracing computation as foundational elements rather than afterthoughts. While traditional statistics education has prioritized theoretical frameworks and applications, computation has emerged as the backbone of contemporary data analysis—from data acquisition and wrangling to visualization, modeling, and communication. Now, AI tools are further transforming this landscape, creating both opportunities and challenges for statistics and data science educators. The presentation will outline a forward-looking curriculum model for introductory courses that balances statistical thinking, data science methods, and explicit computational instruction.
2:30 - 2:50 PM EDT
Data and Computing in K12: implications for undergraduate data science education
11th Floor Lecture Hall
Speaker
Nicholas Horton, Amherst College
Session Chair
Anarina Murillo, Brown University
Abstract
In recent years, data and computing have become an increasingly important part of K-12 education. Early exposure to data literacy, coding, and computational thinking equips students with foundational skills that can accelerate and deepen their engagement. As students arrive at college already familiar with basic data concepts and tools, undergraduate courses and programs will need to shift focus from introductory instruction to more advanced topics. In this talk, I will discuss the landscape of data and computing education in K-12 and implications for data literacy and undergraduate data science.
3:00 - 3:20 PM EDT
"Writing to learn or learning to write: Writing in the statistics curriculum and the role of LLMs "
11th Floor Lecture Hall
Speaker
Alex Reinhart, Carnegie Mellon University
Session Chair
Anarina Murillo, Brown University
Abstract
Statistical writing is a key part of the undergraduate statistics curriculum, both because it helps students reason about complex ideas and because it teaches them an important professional skill. Through writing, students learn to construct arguments about data, present evidence, and communicate results. Large language models (LLMs) thus raise important questions about the role of writing in statistics education: can they serve as useful tools for students still mastering writing, or do they impede development of important skills? We describe recent research comparing the style and structure of student-, expert-, and LLM-written statistical reports, revealing unexpected biases in the writing style of LLMs, and discuss the implications for students still learning to write like experts.
3:30 - 4:00 PM EDT
Coffee Break
11th Floor Collaborative Space
4:00 - 5:00 PM EDT
Group Work
5:00 - 6:30 PM EDT
Reception
11th Floor Collaborative Space
Wednesday, May 21, 2025
9:00 - 9:05 AM EDT
Teaching Parallel Computing and Forecast Analytics through Real-World Data: A Viewership Case Study
Lightning Talks - 11th Floor Lecture Hall
Speaker
Shanzhen Gao, Virginia State University
Session Chair
Alice Paul, Brown University
Abstract
This lightning talk presents a dual-purpose teaching and research framework integrating parallel computing concepts with real-world data analytics. Using daily viewership data from a widely accessed research paper, students use a parallel data entry simulation to explore decomposition, synchronization, and speedup concepts. The dataset is then analyzed using Python and forecasted using Linear Regression, ARIMA, Prophet, and LSTM models. This hands-on activity bridges abstract computing concepts with applied data science and fosters student engagement through visualization, collaborative learning, and predictive modeling. The approach is scalable and reproducible across computer science, business analytics, and STEM curricula.
9:05 - 9:10 AM EDT
Final Exam Scheduling at Bucknell University
Lightning Talks - 11th Floor Lecture Hall
Speaker
Lucas Waddell, Bucknell University
Session Chair
Alice Paul, Brown University
Abstract
This talk will introduce an interdisciplinary undergraduate research project concerning the creation of optimized final exam schedules at Bucknell University. This project, which involves data visualization, mathematical optimization (integer programming), and web-app development, resulted in a tool that has been used by the Bucknell Registrar's office to schedule final exams since the Spring 2024 semester.
9:10 - 9:15 AM EDT
Reflections on a course in dimension reduction
Lightning Talks - 11th Floor Lecture Hall
Speaker
Peter Gao, San Jose State University
Session Chair
Alice Paul, Brown University
Abstract
In this talk, I’ll discuss reflections from teaching a course on dimensionality reduction for data analysis to applied math, statistics, and data science Master’s students. The course ties together material in linear algebra, statistics, and probability and aims to prepare students for future study in statistics and machine learning. In covering such a broad variety of topics, I have strived to focus less on technical details and to emphasize differences in perspective between methods. For example, factor analysis and principal components analysis resemble each other superficially but have different motivations and justifications. Challenges faced in the course include making the material accessible to students coming from a wide variety of backgrounds and ensuring students with limited experience with linear algebra and proof-based math (and students with limited applied data analysis skills) can achieve key learning objectives.
9:15 - 9:20 AM EDT
Computation Across Contexts with Quarto Live
Lightning Talks - 11th Floor Lecture Hall
Speaker
Adam Gilbert, Southern New Hampshire University
Session Chair
Alice Paul, Brown University
Abstract
This lightning talk showcases how Quarto Live can bring interactive applied mathematics and statistics into interdisciplinary classrooms, without tedious installations or finicky setup. We’ll demo a browser-based Bayesian inference activity set in a biology context, designed to show students how observed data update prior beliefs. The talk will also highlight how easily these activities can be built, and deployed via GitHub Pages. Finally, we’ll discuss similar examples that can inject applied math and statistics into discussions from psychology, economics, engineering, and more — lowering barriers for instructors to embed computational thinking across the curriculum.
9:20 - 9:25 AM EDT
Data Detectives: Introducing Data Science Through Curiosity and Critical Thinking
Lightning Talks - 11th Floor Lecture Hall
Speaker
Karianne Bergen, Brown University
Session Chair
Alice Paul, Brown University
Abstract
In this lightning talk, I will share insights from teaching a First Year Seminar, Data Detectives: How to Think Like a Data Scientist, which was co-developed with an undergraduate student. The course was open to students of all disciplines, and explored data literacy, analysis, and ethics through hands-on activities – including conducting polls of campus sock preferences, debating pop culture with data, and learning from institutional staff and researchers engaged with data science.
9:25 - 9:30 AM EDT
Statistics and Data Science in Applied Math Education
Lightning Talks - 11th Floor Lecture Hall
Speaker
Manuchehr Aminian, California State Polytechnic University, Pomona
Session Chair
Alice Paul, Brown University
Abstract
The pressing challenge for many of us at this workshop is fitting three undergraduate degrees into one to give our students vital knowledge and skills for the workforce and being an informed member in society. Compromises have to be made to do this, but an easy way to begin is a strategic choice of application and motivating problems in our courses. As a mathematician, I will discuss my experiences incorporating elements of data analysis and statistics in math courses I've taught; including single and multi-variable calculus, introductory programming for mathematicians, and undergraduate and graduate numerical analysis. I hope to foster a conversation and sharing of ideas in this direction.
9:30 - 9:35 AM EDT
Teaching Simulation and Process Design with Case Studies
Lightning Talks - 11th Floor Lecture Hall
Speaker
Adrian Ramirez-Nafarrate, Tecnologico de Monterrey
Session Chair
Alice Paul, Brown University
Abstract
In this talk, we present the main features of a case study developed to support the teaching of simulation, business process analysis and design, and operations management. The case, titled "The Passport Nightmare," is inspired by the author's personal experience in obtaining a passport for his newborn children. It has been carefully crafted to illustrate key concepts such as inefficiencies arising from scheduling based solely on averages, batch processing, rework, and resource mismanagement. The case also introduces the use of resource pooling, the integration of information systems to streamline processes, the application of Little’s Law, and the role of simulation in designing improved business processes. Additionally, we emphasize the pedagogical value of using custom-developed case studies in analytics education and share key insights for creating a successful learning experience.
9:35 - 9:40 AM EDT
Motivating Data Cleaning with Election Returns
Lightning Talks - 11th Floor Lecture Hall
Speaker
Daryl DeFord, Washington State University
Session Chair
Alice Paul, Brown University
Abstract
In this lighting talk I will discuss an assignment from my introductory data and statistics courses that motivates the importance of using mathematics for data cleaning. Students are provided with synthetic vote returns representing a collection of precincts in Virginia and asked to determine the percentage of all voters who voted for candidate A. This seemingly simple task requires students to identify flaws in the data and formulate internal inconsistency tests between the columns before documenting and justifying their decisions. After submission, this is paired with a classroom discussion about the fact that among a class of 30-50 students there are usually at most a handful of duplicate final answers.
9:40 - 9:45 AM EDT
Trend analysis of snow data with changepoint detection
Lightning Talks - 11th Floor Lecture Hall
Speaker
Jaechoul Lee, Northern Arizona University
Session Chair
Alice Paul, Brown University
Abstract
This presentation introduces methods for estimating long-term trends in two types of snow data: daily snow depth series and weekly snow presence series. The methods are based on Markov Chain models that account for seasonality, support set characteristics, autocorrelation, and changepoints in mean levels. Changepoints—often caused by station relocations or changes in instrumentation—are critical to address when estimating long-term trends. To detect the number and time locations of these changepoints, we employ Genetic Algorithms that optimize a penalized likelihood function based on the minimum description length principle. The methods are demonstrated using a daily snow depth series recorded near Warm Lake, Idaho, and a weekly snow presence series from Snowmass Village, Colorado.
9:45 - 9:50 AM EDT
Posterior Predictive Design for Phase I Clinical Trials
Lightning Talks - 11th Floor Lecture Hall
Speaker
Shouhao Zhou, Penn State College of Medicine
Session Chair
Alice Paul, Brown University
Abstract
Bayesian learning is foundational to adaptive clinical trial design methodology. In this recent work published in JASA, we introduce the Posterior Predictive (PoP) design, a novel interval-based approach that leverages advanced Bayesian predictive hypothesis testing techniques to guide dose transitions in phase I clinical trials. Unlike existing interval-based methods, the PoP design achieves global optimality, ensuring consistent convergence to the true maximum tolerated dose (MTD) at the theoretical rate of 𝑛^{−1/2}. Extensive simulation studies demonstrate that the PoP design substantially improves operating characteristics for MTD identification, offering a significant advancement in practice.
9:50 - 9:55 AM EDT
A Tale of Optimization in Practice
Lightning Talks - 11th Floor Lecture Hall
Speaker
Thiago Serra, University of Iowa
Session Chair
Alice Paul, Brown University
Abstract
We often sell (or are sold on) mathematical optimization as an objective form of representing a system to make it more efficient. If that can work for a small operational unit, such as a production line, why not applying it to an entire organization or to the whole economy at once? This is a question to which Francis Spufford returns at every appearance of the Nobel laureate Leonid Kantorovich as a character in his masterpiece Red Plenty. From reading excerpts of this book through a course about optimization modeling, I invite students to reflect on the challenges associated with implementing large scale projects in organizations with multiple stakeholders.
10:00 - 10:30 AM EDT
Coffee Break
11th Floor Collaborative Space
10:30 - 11:15 AM EDT
Group Work
11:15 AM - 12:15 PM EDT
Large Group Report Out
Group Presentations - 11th Floor Lecture Hall
12:15 - 12:30 PM EDT
Closing Remarks
11th Floor Lecture Hall
Alyssa Bilinski, Brown University
Anarina Murillo, Brown University
Alice Paul, Brown University