Long-Term Forecasting of Hydrocarbon Products Sales

Many industries that use Liquefied Petroleum Gas (LPG) as an energy source are experiencing a steady rise in demand, particularly for GPL/C (LPG for automotive vehicles). This study aims to enhance the prediction of GPL/C sales in Algeria over the next five years by combining both traditional and advanced forecasting techniques. We collected historical sales data from 2014 to 2022 and conducted exploratory data analysis to uncover underlying trends and seasonal patterns. Data preparation techniques were applied to address missing values and manage outliers, ensuring the integrity of the dataset. The SARIMA (Seasonal Autoregressive Integrated Moving Average) method was then utilized to forecast GPL/C sales. We verified the stationarity of the data using the Dickey-Fuller test and applied differencing methods to obtain a stationary time series. The performance of the SARIMA model was assessed through various metrics to evaluate its accuracy and robustness. Our findings indicate a notable upward trend in dGPL/C sales for the upcoming years, accompanied by mild seasonal fluctuations. The SARIMA model proved effective in predicting future GPL/C sales, successfully capturing both trends and seasonal variations in the data.