In brief
- Key staff members in the Westfield State University Office of Institutional Research and Assessment exited the university just as critical state compliance deadlines loomed, exposing deeper gaps in data governance, longitudinal tracking, and analytic capability.
- What began as emergency support evolved into the design and implementation of a full student success analytics platform, complete with eight years of reconstructed historical data, automated reporting pipelines, and a predictive model that identifies at-risk students.
- With the cloud-based platform now in place, the university can meet state and federal reporting requirements with confidence, inform leadership decisions with reliable data, and proactively intervene to improve retention, advance equity, and strengthen institutional effectiveness.
Challenge: Mounting pressure, missing data, and no clear path forward
Westfield State University's Office of Institutional Research and Assessment (OIRA) faced an unprecedented crisis when key personnel exited just as annual state compliance reports were due. With no staff in place to meet mandatory IPEDS and Massachusetts Department of Higher Education submission requirements, the university needed outside expertise.
As the project progressed, a far more complex situation emerged. WSU's census files were overwritten for years, leaving no reliable longitudinal record of student enrollment, demographics, or outcomes. Even simple questions, such as how enrollment in a given department had trended over time, couldn't be answered without reconstructing data from transactional files that rarely matched official reported figures.
A New England Commission of Higher Education (NECHE) accreditation site visit had already flagged data governance and longitudinal tracking as significant gaps in WSU's reporting. The absence of reliable data posed a critical barrier to decision making for the university's newly appointed provost — a challenge made worse by limited technical resources, siloed systems, and internal hesitation about bringing in outside help. The pressure to act was mounting. WSU needed to improve student outcomes to secure continued state financial support, all while navigating growing accountability requirements, enrollment volatility, and constrained resources.
Further, without access to accurate data, it was not possible to assess existing student support programs to determine if continued investment in those programs was warranted.
Approach: Building a lasting analytics foundation from the ground up
What began as emergency OIRA coverage grew into a comprehensive effort to build a lasting student success analytics capability, structured around four connected workstreams:
- Strengthening data governance
- Reconstructing eight years of longitudinal data
- Constructing an analytics platform
- Uncovering actionable insights and aggregating predictive analytics
Strengthening data governance
A student success data governance working group brought together academic affairs, IT, and institutional stakeholders to define KPIs, document source systems, identify data gaps, and develop remediation plans. Running the project through the Office of the Provost as executive sponsor and involving IT from the outset reduced internal resistance and kept momentum without overburdening any single office.
Reconstructing eight years of longitudinal data
Massachusetts institutions are required to submit raw, detailed data files to the State Longitudinal Data System every year. After retrieving WSU's historical submissions from the state, the team reconstructed eight years of census-equivalent student data that had been lost to annual file overwrites, mapping historical submissions against current definitions, normalizing data elements, and delivering the university’s first comprehensive longitudinal dataset.
Constructing an analytics platform
The team built a cloud-based student success analytics platform on Amazon Web Services (AWS), with automated extract-transform-load pipelines that make reporting repeatable and reliable. Dashboards were designed in close collaboration with academic affairs to surface the metrics WSU needed most, including state-required on-time credit accumulation and a high-enrollment, low-pass-rate course indicator that flags early warning signals for retention and completion.
Uncovering predictive analytics and actionable insights
Beyond compliance reporting, the team built an AI-enabled predictive at-risk model that draws on variables including high school GPA, financial aid status, registration behavior, and distance from campus to generate a return-probability score for every student. As new data arrives each year, the model automatically adjusts its own variable weightings, staying current without ongoing data science staffing.
To enable easy access to insights, Huron built a conversational AI chat agent within the AWS environment. An AI chatbot was uniquely suited for the challenge at hand. Rather than navigating dashboards or filters or requiring additional visualizations, users can ask a question as they would a colleague, and the agent interprets intent, retrieves the right data, and returns a clear answer. Rather than waiting on IT or learning specialized query tools, users can ask questions in plain English. For example, they can probe the differences in financial profiles between at-risk and not-at-risk students and receive a specific, data-grounded answer in seconds. This natural language capability transforms data from a resource only specialists could access into a tool every user can rely on in real time, right when a decision needs to be made.
Throughout the engagement, the team prioritized building internal capability over creating dependency on external resources. All underlying data remains accessible and exportable, and structured training and handoffs were built into every phase.
Results: From compliance crisis to confident, data-driven leadership
Compliance confidence
WSU now consistently and accurately produces mandatory student success metrics. Compliance reporting has moved from a manual, often error-prone undertaking to a reliable institutional capability grounded in clean data and automated processes.
Historic longitudinal data at their fingertips
For the first time, WSU leadership has a continuous, reliable view of historical student data. The provost and academic leaders can answer foundational questions about departmental enrollment trends and program performance over time, and confidently make planning and resource decisions.
Actionable student-level insights
The predictive at-risk model, which reached approximately 80% confidence after initial delivery, gives advisors an early-warning system of real precision. Staff can use the student ID to identify students flagged as high-risk at the start of a term, along with the factors driving that risk. One standout finding: Students with more than $5,000 in unmet financial need face a substantially elevated risk of not returning, a data point that enables more targeted deployment of aid and advising resources.
Strengthened institutional effectiveness
The data governance structures and longitudinal reporting built through this engagement directly support WSU's accreditation posture and overall institutional effectiveness and have sparked genuine enthusiasm among faculty and staff for what the data can reveal. Assessment of existing programs, some of which have been in operation for decades, is now being conducted, providing opportunities to revise ineffective programs or shift those resources to more effective ones. This information will lead to significant improvements in student success.
A foundation for future growth
The platform is designed to grow. Future enhancements under discussion include National Student Clearinghouse integration, licensure pass-rate tracking, graduation and course-completion prediction models, and admissions-stage enrollment forecasting.
“This endeavor has given us the data we need to assess the effectiveness of existing student support efforts. For example, our LEADS program is more than 50 years old and, until recently, we didn’t have data to assess if it’s benefiting the participating students. Our new data suggest that it isn’t effective, allowing me to make changes in the program."
-Dr. William Salka Provost and Vice President for Academic Affairs, Westfield State University
What began as an emergency triage became a scalable analytics foundation. In just one year, WSU moved from disorganized, missing data to longitudinal insight and predictive capability, gaining the governance, data, and tools to operate as a truly data-informed, student-centered institution.