Introduction
BDA, which refers to big data analytics, is receiving much more attention in the supply chain management field. The reason for this is that BDA has a variety of applications in the mentioned field, which include trend analysis, customer behavior analysis, and demand prediction (Seyedan & Mafakheri, 2020). The paper reports on a survey that was conducted to examine the predictive BDA applications in supply chain demand forecasting to suggest a categorization of the applications. The authors have classified the algorithms and the applications in SCM into time time-series forecasting, K-nearest neighbors, clustering, neural networks, support vector regression, regression analysis as well as support vector machines. They also point to the idea that the literature is specifically insufficient on the applications of big data analytics: in the event of closed-loop supply chains and accordingly point out the avenues for future study.
Demand forecasting is an area in predictive analytics that attempts to comprehend as well as foresee consumer demand to optimize supply decisions by corporate supply chain plus business management. The methods used in this procedure are broken into qualitative and quantitative. The two methods are based on experts’ opinions and historical sales information. This procedure can be utilized in inventory management, production planning, as well as assessment of future capacity requirements.
Research Conducted
This is a survey that aims to examine predictive big data analytics: in supply chain demand forecasting to recommend or suggest a categorization of the applications. The survey is conducted to help identify the gaps in the available literature as well as offer knowledge for future studies.
Journal Background & Literature Survey
Currently, businesses are adopting ever-increasing accuracy promotion efforts to maintain competitiveness plus grow or maintain their profit margin. Forecasting models have been utilized in accuracy promotion to identify as well as fulfill the customer’s needs plus preferences using predictions obtained from customer information as well as transaction records to manage products SC accordingly. Demand forecasting refers to predicting the demand of materials to ensure the right products are delivered in the ideal quantity to fulfill customer demands without creating a leftover. Errors in forecasting can lead to leftover or excess, which is not only wasteful but also costly (Seyedan & Mafakheri, 2020). The process is essential to the supply chain since it provides helps in developing operational strategies. It is an underlying hypothesis for strategic business processes plus the start point for the majority of the supply chain activities such as purchasing, raw material planning, inbound logistics, and manufacturing. Demand forecasting also enables critical business tasks such as financial planning, risks assessment, production planning, plus the purchase of raw materials.
Forecast accuracy enables retailers to avoid overstocking and stock-outs, better the production lead times, reduce costs, improve operational efficiency as well as better the customer experience. It is essential to understand that sales forecasting is more than utilizing sales data from before to establish customer demand in the present or future (Seyedan & Mafakheri, 2020). The procedure can be divided into quantitative and qualitative forecasting, both dependent on various resources as well as data sets to infer valuable sales data. The former method is utilized in the case of available historical sales information on particular items plus a pre-determined demand. It needs mathematical formulae as well as data sets such as sales, fiscal reports, website analytics, and revenue figures (Nunes et al., 2020). Qualitative forecasting depends on new technologies, product lifecycle, pricing and availability changes, product upgrades, and the experience and intuition of individuals responsible for forecast planning.
Findings and Conclusions
A key finding from the review of the available literature is that there is a very preliminary study done on the big data analytics applications in reverse logistics and CLSC. There are benefits as a result of adopting a knowledge-driven strategy for design plus management of CLSCs. Because of the increasing awareness of the environment as well as incentives from governments, a tremendous amount of returned items are collected, received, and organized in various points of collection. The uncertainties have a direct influence on the cost-efficiency of manufacturing procedures, the last price of the refurbished products, plus their demand. The design, as well as operation of CLSCs, brings out a case for BDA from both the demand as well as supply forecasting standpoints.
The increasing need for analyzing consumer behavior as well as demand prediction is fueled by globalization plus a rise in market competition and the surge in supply chain digitization practices. The authors conducted a detailed review of the big data analytics applications in supply chain demand prediction in the study. They overviewed the big data analytics methods used and offered a comparative classification. They collected and evaluated the available studies regarding the procedures and mechanisms utilized in demand forecasting. Several mainstream mechanisms were found and researched with their advantages and disadvantages. The NN (neural networks), as well as regression analysis, are viewed as the two mainly used mechanisms (Law et al., 2019). The analysis also discovered that optimization models could be applied in improving forecasting precision via formulating as well as optimizing a cost function for the fitting of the forecasts to data.
Summary of Learning
The journal enables a reader to understand that the data available in the supply chains contain valuable information. Evaluation of such information can allow forecast trends of markets, customer behavior, as well as prices. This can be helpful to organizations as it can facilitate their adapting to competitive surroundings (Nunes et al., 2020). To predict demand in a supply chain, various predictive big data analytics algorithms have to be applied. The algorithms could offer predictive analytics using time-series strategies, associative forecasting methods as well as auto-regressive methods. The forecasts from the algorithms could be incorporated with item design attributes and online search traffic mapping to combine both price and customer information.
The journal also provides knowledge concerning forecast accuracy, which allows a retailer to avoid issues of overstocking and stock-outs. It also helps them better their production lead times, reduces costs, improves operational efficiency as well as better the consumer experience (Law et al., 2019). It shows that it is critical to understand that sales prediction is more than utilizing sales data from before to decide on consumer demand in the current time or future (Nunes et al., 2020). The authors also relay that the procedure of demand forecasting can be grouped into quantitative and qualitative forecasting, both being dependent on various resources as well as data sets to infer valuable sales data. The former method is utilized in the case of available historical sales information on particular items plus a pre-determined demand. At the same time, the latter is dependent on new technologies and product lifecycle.
References
Law, R., Li, G., Fong, D. K. C., & Han, X. (2019). Tourism demand forecasting: A deep learning approach.Annals of Tourism Research, 75, 410–423. Web.
Nunes, L., Causer, T., & Ciolkosz, D. (2020). Biomass for energy: A review on supply chain management models. Renewable and Sustainable Energy Reviews, 120, 109658. Web.
Seyedan, M., & Mafakheri, F. (2020). Predictive big data analytics for supply chain demand forecasting: Methods, applications, and research opportunities. Journal of Big Data, 7(1). Web.