Faculty Spotlight: Joshua Loyal

Joshua Loyal is an assistant professor in Florida State University’s Department of Statistics and a faculty adviser for students in the Interdisciplinary Data Science Master’s Degree Program statistics concentration, both of which are within the College of Arts and Sciences. His research focuses on statistical network analysis and using models to understand connections within systems, such as the system created when, for example, friend A introduces friend B to friend C. After Loyal earned dual bachelor’s degrees in mathematics and statistics in 2013 from Duke University, North Carolina, he received his master’s in physics in 2016 from Yale University in Connecticut. From 2015 to 2018 he worked at DataRobot, Inc. as a data scientist. He then earned a doctorate in statistics from the University of Illinois Urbana-Champaign in 2022 before joining FSU’s faculty.

Tell us a little about your background, where you’re from and what brought you to FSU.

My dad was an FSU fan, so I was very aware of the university. The well-established nature, strong leadership, and exciting collaborations within the statistics department were all driving factors when deciding where to take my career. I was interested in statistics because I wanted to understand how to use statistical mathematics to draw meaning from data. I wanted to develop new statistical methodologies, which you can also do in industry, but I like the flexibilities available in academia.

Break down your research for us.

My primary area of research is statistical network analysis. This analysis can be applied in examining how and why people’s interactions with social media content evolves over time. Statistical network analysis uses statistical tools to give people rigorous insights into the mechanisms driving the interactions between the entities in their data. Network data is everywhere — we can also use these analyses, for example, to learn more about Alzheimer’s disease. Scientists can study brain networks to see how brain connections change over time and identify potential correlating factors that exacerbate the condition, such as lifestyle choices or genetic traits.

I also develop new statistical models for dynamic networks, or systems that change over time, such as the evolving trade relations between nations. My work provides people new ways to understand and visualize changing behaviors and connections among entities by using a statistical tool called Bayesian analysis. Dynamic networks are used in many applications in sociology, public health, bioinformatics, and neuroscience, to name a few.

What makes you passionate about your research?

While pursuing my doctorate, I worked on a project where we looked at how alliances in online video games changed over time. I noticed the statistical tools used to analyze the alliances lacked a lot of rigor, which could lead to incorrect conclusions. This experience opened my eyes to a field that has many applications but needed more work to generate tools that provide accurate, reliable insights.

What inspired you to choose your field?

I’m interested in using math to understand real-world problems. While working in experimental high-energy particle physics during my undergraduate and master’s programs, I understood statistics was the underlying framework used to gain insights from data with mathematics. The mathematical and application aspects of statistics are very interesting to me.

What do you want the public to know about the importance of your research?

It’s important to use the appropriate tools so we can achieve confident results. The tools I’m developing that detect patterns and therefore potential causes of problems allow us to make conclusions with confidence. These developments give scientists more agency to answer important questions.

What’s your favorite part of your job?

I enjoy being able to think deeply and thoroughly about new problems and work toward solutions. I also enjoy working with students to explain fundamental concepts. It’s easy to get lost in the weeds, but I like setting up guiding paths for students to understand concepts and develop the intuition they need to be successful researchers, data scientists, or network experts in different industries.

Tell me about being a faculty adviser for IDS students.

I guide and mentor students concentrating in statistics based on where they want to use their degree and match them with electives that help them achieve their career, professional and personal goals.

What’s your best memory from working at FSU?

Aside from research, it’s always great to see others’ hard work pay off, such as hearing about other faculty members’ research in statistics and seeing my students pass their essay defenses and develop as independent researchers.

Tell us about some upcoming projects or goals.

Assistant professor of statistics Jonathan Stewart and I recently received a $300,000 National Science Foundation grant, which’s exciting. Our project examines how relationships are interconnected and how new connections often emerge through existing ones. We’re developing model frameworks that better reflect how connections influence one another. If I have a friend, for example, they may introduce me to a new friend; this situation and my network depends on the connections I have. We are developing a model that captures this dependence. Our project’s also very interesting as we’re working on new models incorporating existing modeling methods but now featuring a dependent structure.

Additionally, there’s an issue in our field regarding how statistical network analysis applications have trouble processing massive real-world networks like growing social media networks with millions of users. My students and I are working to develop tools to overcome this problem and analyze large networks. We want to know the certainty of our answers, so these projects provide tools with rigorous uncertainty quantification to provide accurate answers for large systems.

If your students only learned one thing from you (of course, hopefully they learn much more), what would you hope it to be?

A good researcher and data scientist is someone who’s curious and critical toward data. When given data, make sure to look at it, and think deeply about what tools are appropriate for the analysis question. It’s also important when applying a statistical method to look for any shortcomings so you can be confident in your conclusions.