Expense Anomaly Index: Uncovering Hidden Signals in Spending

University institutions today are under even greater scrutiny from state regulators and auditors with increased pressure to ensure the transparent and compliant use of funds in higher education. However, finance divisions lack an automated method to identify spending anomalies early in the process. This often results in delayed corrective actions, reactive audit responses, and potential reputational and financial risks. This project introduces an unsupervised machine learning approach of Local Outlier Factor (LOF) and Isolation Forest (IF) to address this gap by developing an anomaly detection index for expense audits. Utilizing these results, finance divisions can develop an internal benchmark on a [0-1] scale and begin to monitor these Anomaly Indices via a Tableau dashboard. This addresses the gap by equipping institutions with tools to proactively monitor and mitigate risks