Florida State University Builds for the Future with AI and Self-Service Analytics

Selecting a university is a big decision. As higher education hopefuls weigh their options, many factors influence the decision — from professional goals and financial constraints to a university’s curricula, programs and prestige. Consistently ranked among the top public colleges and universities in the United States, Florida State University (FSU) attracts thousands of new students each year with its highly rated academic and athletic programs, reasonable cost of attendance, and strong reputation as a preeminent research university.

Data is essential to FSU’s strategic objective of operational excellence. With Snowflake, FSU has built a solid data foundation for powering analytics and AI use cases that help the institution advance its mission.

Self-service analytics isn’t a new concept at FSU. Teams have been relying on data and analytics for at least two decades. But the university’s previous data architecture was difficult to scale and maintain. Integrating additional data sources was burdensome and compounded FSU’s existing data storage constraints, while delivering new data products required 90 days or more, which frustrated end users. 

Seeking to modernize the university’s data infrastructure, FSU’s data and analytics team began planning a migration to Snowflake. Around the same time, FSU’s Campus Reimagined program — a cross-functional effort to prepare for the class of 2030 — had already adopted Snowflake due to its native SQL support, ease of use and ecosystem of technology partners. Success with Campus Reimagined laid the groundwork for a healthier data culture, greater data literacy and a growing community of Snowflake advocates. As a result, Snowflake became a go-to solution for accelerating data innovation across the institution.

Empowering faculty and staff with Snowflake elevates productivity for FSU’s data and analytics team. “Getting away from the daily grind of building reports and answering questions allows the team to focus on strategic initiatives,” says Jorge Vidal, FSU’s Senior Director of Data and Analytics. Snowflake Secure Data Sharing and built-in data governance and security features through Snowflake Horizon Catalog help FSU expand data access while maintaining proper controls.

Today, FSU stores data from 47 discrete sources in Snowflake — an increase of more than 11x compared to its prior data platform, Oracle. Onboarding data sources no longer requires weeks of work for technical staff. Snowflake’s elastic performance engine with optimized storage helps the institution analyze more data and overcome bottlenecks. For example, one of FSU’s queries that previously required four hours to run now takes just 10 seconds with Snowflake.

With Snowflake, FSU has achieved a secure, scalable process for democratizing data. Since much of the institution’s data is already stored in Snowflake, staff simply submit a request form to gain access within hours. Connecting and sharing new data sets takes about five days — instead of 90 — with minimal data engineering. According to Vidal, “Copying a data set from point A to point B on Snowflake is extremely simple and fast.”

Self-service data and analytics make it easier to analyze FSU’s facility usage and occupancy, applicant yield and scheduling variables. Increased access to reliable data via Snowflake fuels decentralized innovation with less help from IT. “Now people actually generate their own information products and share them,” Vidal says. 

FSU’s Office of Institutional Research relies on Snowflake to respond faster to stakeholders’ needs and understand data lineage. “Snowflake has given FSU’s Institutional Research team the ability to combine otherwise disparate data sources for greater sophistication in our analytics efforts, and more than ever before we are able to deliver high-quality data dashboards and other tools that empower our university to improve the success of students, faculty and staff,” says Dr. James Hunt, Director of Institutional Research at Florida State University.

Snowflake’s Document AI provides an easy-to-use way to capture unstructured data from FSU’s data request forms. Extracting the request reason, date, approver name and other information from PDFs streamlines data auditing and outlier detection while bolstering security and privacy. FSU also uses Document AI to efficiently extract unstructured data from third-party sources. 

FSU plans to use Document AI to an even greater extent in the future, especially for healthcare-related projects. For example, Vidal believes Document AI could unlock diagnosis insights from unstructured data trapped in physicians’ patient notes. “I’m proposing to take all these notes we’ve captured in the past and turn that into quantifiable numeric data,” Vidal says. “This would take weeks and weeks to do manually. But with Document AI, we could very quickly surface valuable insights, such as how many patients with a diabetes diagnosis also have cardiovascular disease, dementia and so on.” 

Rearchitecting on Snowflake helps FSU navigate challenging situations without losing sight of the student experience. During the COVID-19 pandemic, FSU used data in Snowflake to monitor remote student engagement, track population infection patterns, inform classroom capacity decisions and override parking violations for people in isolation. Accessing case count and other COVID-19 data via Snowflake Marketplace saved FSU many hours of development and maintenance work. According to Bucior, “COVID drew a big, bold underline beneath Snowflake’s value proposition.”

FSU’s space utilization dashboard brings together class enrollment, wireless network usage and card swipe data to support better resource usage for the university and smarter class scheduling for students. “Inefficient use of our spaces can lead to unnecessary construction, lighting, heating, cooling and so on,” Vidal says.  “This dashboard helps us save costs, improve efficiencies and further our green agenda — and it’s all made possible because we have access to these different data sources in Snowflake,” Vidal says.

With Snowflake, FSU is well-positioned to respond to the institution’s evolving data needs. “AI is impacting the university, and we all know the importance of having high-quality data to build your AI models,” Vidal says. “Snowflake is vital as we look to bring AI to our data in the next phase of our data journey.” With Snowflake as its data foundation across departments, FSU will further increase accuracy for modeling and forecasting while also migrating remaining reporting use cases to reduce complexity and cost. 

FSU also plans to build a data vault to continue to improve efficiency and transparency. Vidal says, “Snowflake’s performance is going to allow us to build a highly extensible and performant data vault that will go far beyond replicating tables to save us money while powering our evolving data needs for ongoing success.”