Time Series Modeling and Forecasting of Petrol Sales at a Fuel Station in Zimbabwe.

Abstract:

Petrol is an important commodity in the Zimbabwean economy. Its availability, shortage and pricing have a significant impact on key economic indicators and the livelihoods of the general population in the country. This study focuses on petrol sales volumes at a fuel station in Zimbabwe. There are costs and negative implications that emanate from overstocking or understocking of petrol at the fuel station. This makes it necessary to forecast future demand, so as to plan and maintain adequate volumes of petrol. This study sought to fit a SARIMA model using monthly sales data from January 2015 to December 2024. Minimal information criteria were used to compare and select from candidate models. Parameter p-values were then used to identify and eliminate likely insignificant parameters. This was because insignificant parameters contribute to model overfitting. The best fitting model was then used to forecast petrol sales volumes for the months from January to December 2025.