Examining annual and seasonal variations in turnover within the food production sector: a case study of Simbisa Brands (Steers)
- Author
- Hove, Grace
- Title
- Examining annual and seasonal variations in turnover within the food production sector: a case study of Simbisa Brands (Steers)
- Abstract
- This study focuses on investigating the seasonal and annual turnover variations of Simbisa Brands (Steers) in Zimbabwe from 2020 to 2024. This was achieved by employing advanced time series forecasting models to gain insights on the turnover dynamics. In particular, this research study compares the predictive performance of the Seasonal Autoregressive Integrated Moving Average (SARIMA) model and machine learning-based Extreme Gradient Boosting (XGBoost) algorithm. Several diagnostic tests were done to validate data reliability. Models were developed to Multiple accuracy metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R^2) were used for model evaluation. These accuracy metrics revealed that XGBoost significantly outperforms SARIMA in capturing complex, nonlinear, and seasonal turnover patterns. From the model results, XGBoost had a MAPE of approximately 15.49%, an RMSE of 0.568, and an R^2 of 67.94%, compared to SARIMA’s 102.48%, 1.164, and 6.7% respectively. Graphical comparisons revealed that XGBoost closely follows turnover fluctuations, including peak and trough periods, whereas SARIMA’s linear assumptions limit its predictive accuracy especially during structural breaks and sudden changes. The findings of this study suggest that machine learning models like XGBoost are better suited for dynamic, complex environments, thereby providing more reliable forecasts to inform production planning, inventory management, and strategic marketing decisions. This research advocates adopting to data-driven forecasting tools within the Zimbabwean fast-food industry. Overall, the study reinforces the transformative potential of integrating machine learning models into operational frameworks to optimize performance in the face of market volatility.
- Date
- June 2025
- Publisher
- BUSE
- Keywords
- seasonal and annual turnover
- Supervisor
- Ms. J C Pagan’a
- Media
-
GRACE HOVE - SFM.pdf