IRSA: Understanding Science Through Statistics

Christina Knudson standing in front of a presentation on Bayesian statistics.
Christina Knudson presents on Bayesian statistics at the noRth conference in August 2019.

“Statistics is the science of learning from data,” says the opening line of the Institute for Research in Statistics and its Applications (IRSA) website. Although that definition might seem broad for a subject that students can pursue as an undergraduate student or PhD candidate, it clarifies the fundamental role of statistics in the even more expansive field of data science. IRSA’s programs teach the campus and community that while the collection of data is important, the ability to analyze, interpret, and implement changes based on the data is far more valuable.

Collaborative Statistics

When it began in the fall of 2016, IRSA originally set out to highlight the importance of statistical reasoning and its applications across a wide range of professional fields. The term “data science” was gaining popularity, and leaders at the School of Statistics recognized the importance of their discipline as the bedrock of this field. As Ansu Chatterjee, IRSA’s director, states, “We had been thinking for quite a while that instead of isolated efforts, we need a more structured platform to study and work on various aspects of data science, including its foundations and theory, methodological tools, algorithms and software, and application in various domains.” To that end, one of IRSA’s focuses is planning educational events—like workshops, short courses, and conferences—in order to foster interdisciplinary collaboration, constantly searching for new ways to help others understand the value of statistics and get the most insight out of their data. 

Group photo of the presenters for the 2019 IRSA Conference

One product of these efforts is their annual conference, which is held at the culmination of the University’s academic year in May. Lan Liu, an assistant professor in the School of Statistics, was the 2019 IRSA conference planning chair, as her main research area coincided with the conference theme about causal inference. This year’s conference was a huge success. As Liu explained, “We had a full audience with a diverse background: epidemiology, statistics, biostatistics, economics, medicine, psychology, and philosophy.” 

The success of this conference is echoed by event participants who provided feedback, stating “the conference contained a great overview of the history of causal inference and fascinating new developments in the field.” Feedback on what was most valuable about the conference noted “the high quality presentations and the cutting-edge material covered,” as well as “incredible speakers from across the country.” The 2019 conference sold out, exceeding the venue’s capacity and doubling registration from 2018 to 2019. This annual IRSA conference fills a vital need for interdisciplinary learning about the application of statistics.

Summer Camps and Student Engagements

IRSA also had a busy summer putting on workshops and camps for students, faculty, and youth in the community. Laurie Derechin, the executive director of the Minnesota Center for Financial and Actuarial Mathematics (MCFAM), says that “MCFAM and the School of Statistics have a long-standing relationship. We have begun working more closely with IRSA and are looking for ways to collaborate on projects of shared interests.” 

Recently, they cosponsored the Machine Learning Summer Camp, a day camp intended to enlighten young students to career possibilities in the STEM fields and to develop machine learning skills. IRSA provided marketing and fundraising. It drew attention to the event’s honorable goals and garnered funding, thus keeping enrollment fees low and providing educational opportunities to those who would otherwise not have the means to participate. 

This past summer, there were two Machine Learning Summer Camps. One was designed specifically to introduce high school-aged women to the world of data science, while the other was open to all high school students interested in the field. Over 40 students attended. “The machine learning projects that students worked on were varied and ranged from gender classification using tweets to handwritten digit recognition,” explained Derechin.

Lindsey Dietz, a senior quantitative analyst for the Federal Reserve Bank of Minneapolis, had a similar idea of engaging the community with the work being done in the statistics field. Dietz received her PhD in statistics from the U in 2016, so when she began exploring the idea of creating a statistical programming conference, she knew IRSA would be the perfect group to collaborate with. Thus, the noRth conference was born. This inaugural conference, held in August 2019, focused on R, a popular coding language in statistics. IRSA cohosted the conference and provided funding, planning, organization, marketing support, and a venue. 

 

conference attendees seated around circular tables working on laptops

Attendees included academics (students and faculty) and professionals from all over the country working in local nonprofit and for-profit industries. This conference provided participants with opportunities to connect and network with other R users and established IRSA and the U as proud supporters of the Twin Cities data science community. As Dietz says, “We received positive feedback about our emphasis on inclusion including a code of conduct and equitable gender representation in speakers.” Dietz is a supporter of IRSA and is glad that noRth was able to expand the influence of the organization. 

Another popular R-focused course hosted by IRSA, Introduction to Reproducible Data Science in R, was back for its third consecutive summer. Alicia Hofelich Mohr, a research support coordinator at the Liberal Arts Technologies and Innovation Services (LATIS), has co-taught this course for the past three years. As she explains it, “This course is targeted towards people who are interested in learning how to analyze and visualize data in R but do not have any experience with statistics or using R.” Participants in this course include both students and staff as well as community members. 

Participants love the interactive hands-on learning opportunity that provides them with the chance to work through exercises, play with data, troubleshoot with instructors, and ask questions. The feedback has been incredibly positive, with some participants stating, “I loved that we were able to start practicing using R immediately,” and “I thought the guided exercises were a great way to apply our learning. The course overall was fantastic and very well organized.” Due to the great success of this program, IRSA now offers a follow-up course, Performing Linear Regression with R. 

Engaging Beyond the Classroom

IRSA has been an asset to the University of Minnesota and the Twin Cities, connecting people with the tools and resources needed to develop statistical reasoning and data science skills. It has helped both students and professors grow professionally and academically. “[IRSA] facilitates communications in research and practical problems within and outside of the School of Statistics,” explains Liu. “Not only does it expose professors to new challenges and networking opportunities, [but it also] helps local businesses better utilize statistical strategies.”  

IRSA is always looking to tackle new problems and further educate people about the value of statistics. By providing learning opportunities for students and the greater Minneapolis community, IRSA can help address the most important questions about data. In Chatterjee’s own words, “We look forward to working with researchers from any discipline where our contribution may add to the state of knowledge.” 

Are you interested in working with IRSA? Email IRSAworkshop@umn.edu with your ideas or propose an event. View upcoming short courses, workshops, and conferences.

This story was written by an undergraduate student in CLA.

Share on: