Polynomial Regression Budget Prediction for Minahasa Fire Department

Abstract

General Background: Public sector budget planning requires data-driven approaches to support service sustainability, particularly for emergency institutions such as regional fire departments. Specific Background: The Minahasa Regency Fire Department budget planning has relied on historical allocations that exhibit fluctuating and non-linear trends across expenditure categories. Knowledge Gap: Previous planning practices have not fully utilized non-linear statistical modeling to interpret historical budget dynamics and medium-term projections. Aims: This study aims to analyze historical budget trends and predict expenditure needs for the 2025–2027 period using a second-order polynomial regression model. Results: Using budget data from 2019–2024, the model successfully captured non-linear expenditure patterns, including a sharp increase in personnel and capital expenditures in 2024, with coefficients of determination of 0.448 for personnel expenditures and 0.409 for capital expenditures, while total projected expenditures approach IDR six billion by 2027. Novelty: This research integrates second-order polynomial regression with an interactive Streamlit-based application for visualizing and interpreting public sector budget projections. Implications: The findings provide an analytical basis for local governments to anticipate medium-term budget requirements, prioritize expenditures, and strengthen data-based financial planning for fire service operations. Highlights • Second-order polynomial modeling captures non-linear expenditure patterns across budget categories• Historical personnel expenditure spikes shape medium-term fiscal projections• Streamlit-based visualization supports data-driven public budget planning Keywords Polynomial Regression; Budget Prediction; Fire Department; Historical Data; Streamlit Application