Oil Production Monitoring Optimization through Predictive Dashboard with Machine Learning Modeling

The demand for oil continues to rise as the population grows, while oil production tends to decline. This creates a challenge for the oil industry in effectively monitoring oil production. Currently, the monitoring process still relies on manual methods, such as Excel, which has contributed to the decline in oil production. To address this issue, this research develops a software-based dashboard integrated with Auto Regressive Integrated Moving Average (ARIMA) and Artificial Neural Network (ANN) models to predict oil production precisely. An analysis of oil well production data from January 2021 to September 2023 shows that ANN outperforms ARIMA, achieving a Mean Absolute Percentage Error (MAPE) of 2.42% and a Root Mean Square Error (RMSE) of 4.52 at a training data split ratio of 80:20. In comparison, the ARIMA model achieved a MAPE of 3% and an RMSE of 4.55 at a training data split ratio of 70:30. A software-based dashboard was developed to visualize the computational and analytical results of the ARIMA and ANN models. This dashboard helps identify wells experiencing production decline, enabling more informed and effective decision-making.