Case Report
Lishura Chen
Abstract
A comprehensive method of finding seasonal patterns of demand and the accurate prediction of future demands are still critical elements for different industries especially in manufacturing companies as it contributes to effective planning and operation. In this paper, a statistical forecasting model has been proposed and implemented in a motorcycle accessories manufacturing company in the USA. Dataset for a 7-year timeframe of historical sales data has been mined, cleaned, and compiled using Python programming. Results have been compared to the conventional forecasting model used by the company. Based on this comparison, using the proposed statistical forecasting model can improve the Mean Absolute Deviation (MAD) by almost 61% and the mean squared error (MSE) by 82%. These improvements will drastically improve the chance of consistently maintaining the right levels of inventory in the right place and at the right time. It also provides the opportunity of ensuring the safety stock of its inventory is sized correctly to avoid inflated carrying costs and lost sales orders due to stock outs.