Forecasting demand for logging and drilling tools is crucial in the oilfield service industry due to their highly volatile demand patterns and complex applications. These tools necessitate substantial capital investment and careful inventory management of materials and supplies (M&S). This capstone project develops a data-driven demand forecasting model to enhance the accuracy of capital expenditure (CAPEX) and M&S planning for the oilfield service industry. By leveraging historical demand data, the methodology combines advanced time series forecasting models, including temporal convolutional network (TCN) and exponential smoothing, to produce forecast results with an accuracy error of 7% over a 6-month horizon, compared to 25% of the current process used by the sponsor company. Adopting the forecast models will lead to more accurate CAPEX budgets and reduced M&S costs, enhancing operational efficiency and resource utilization for the sponsoring company.
