Big data analytics in supply chain management: A state-of-the-art literature review

TL;DR: A novel classification framework is proposed that provides a full picture of current literature on where and how BDA has been applied within the SCM context and reveals a number of research gaps, which leads to future research directions.

About: This article is published in Computers & Operations Research.The article was published on 2017-07-08 and is currently open access. It has received 329 citations till now. The article focuses on the topics: Analytics & Big data.

Summary

Introduction

• As a service to their customers the authors are providing this early version of the manuscript.
• Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

4. Results and discussion

• BDA has been widely adopted to facilitate the supplier selection process and recent efforts have been made to integrate this activity with order allocation problems and to reduce sourcing costs (Kuo et al., 2015).
• In the trend analysis, the results show that prescriptive analytics is the most common and fastest growing area in the BDA-driven SCM, which is closely followed by predictive analytics, while descriptive analytics is receiving less consideration.
• On the other hand, the study of real-time optimization appears to be quite mature in the manufacturing domain with the use of modelling and simulation to develop a real-time production control system, based on streamline context-aware data, generated from tracking devices such as RFID (Babiceanu and Seker, 2016; Kumar et al., 2016).
• Finally, ARM can also be used in the hybrid optimization problems of supplier selection and order allocation (Kuo et al., 2015).

5. Future direction

• AC CE PT ED M AN US CR IP T Indeed, alignment dissolves the boundary across functions.
• Moreover, although the literature has extensively adopted visualization techniques as supplement techniques to predictive and prescriptive models, little attention has been paid on improving data visualization techniques.
• It is anticipated that the framework and operational mechanism of today‘s smart factory would be scalable to the entire SC.
• Conclusion Based on the content analysis methodology of Mayring (2008), this literature review examined 88 journal papers to provide a full picture of where and how BDA has been applied within the SCM context.

Figures

Table 2: Summary of literature by SC function

Table 3: Level of analytics in each SC function

Figure 2: Distribution of reference papers by year

Frequently asked questions

Q1. What are the future works in this paper?

The findings discussed above suggest some future directions to capitalize the research development of BDA applications in the SCM context. To facilitate the horizontal integration throughout the SC, future BDA research should focus more on cross-functional problems such as vehicle routing and facility location, supplier selection and order allocation, demand-driven storage assignment and order picking. To catalyse the rapid progression of BDA application in SCM, future research should balance the focus across all three levels of analytics. Future research should call for this gap because visualizing complex BD would expedite decision making. 

Q2. What are the main applications of BDA in demand management and procurement?

Accurate demand forecasting and early detection of various sourcing risks are among the foremost applications of BDA-enabled predictive models in these two areas. 

Q3. What is the common type of analytics used in demand management?

predictive analytics is the most often common type used in demand management (6 out of 12 papers) and procurement (4 out of 10AC CEPT EDM ANUS CRIP Tpapers). 

Q4. What is the prevalent area of BDA in SCM?

Among five areas of SCM, logistics/transportation (25 out of 88 papers, 28%) is the most prevalent area where BDA is used to support decision-making. 

Q5. What is the importance of data quality and level of trust?

Veracity stresses the importance of data quality and level of trust due to the concern that many data sources (e.g. social networking sites) inherently contain a certain degree of uncertainty and unreliability (Gandomi and Haider, 2015; IBM, 2012; White, 2012). 

Q6. What is the common and fastest growing area in the BDA-driven SCM?

In the trend analysis, the results show that prescriptive analytics is the most common and fastest growing area in the BDA-driven SCM, which is closely followed by predictive analytics, while descriptive analytics is receiving less consideration. 

Q7. What is the main limitation of this paper?

One of the limitations of this paper is that the categorization in the classification framework remains interpretative, as this could lead to concern on subjective bias. 

Q8. What is the main reason why a CLSC is difficult to manage?

Managing a CLSC has always been challenging due to the uncertainties and possible conflicting goals, i.e. profit vs. environment vs. social wellbeing. 

Q9. What is the common approach at the predictive analytics level?

Classification is the most common approach at the predictive analytics level and has been widely applied in manufacturing to support production planning and control (Chien et al., 2014; Wang and Zhang, 2016) and equipment maintenance and diagnosis (Kumar et al., 2016; Shu et al., 2016; Wang et al., 2016). 

Q10. What is the process of the review?

In particular, the review is systemically conducted in accordance with the four-step iterative process:AC CEPT EDM ANUS CRIP T- Step 1: Material collection, which entails a structured process of search anddelimitation of articles.- 

Q11. What is the importance of studying BDA with a holistic perspective?

the review shows that recent research has increasingly recognized the importance of studying BDA with a holistic perspective mindful of SCM as a multilevel, inextricably interlinked system. 

Q12. What is the final layer of the BDA analysis?

The final layer shows the BDA techniques, which can be adopted from multiple data analytics disciplines such as data miming, machine learning, etc.AC CEPTEDM ANUS CRIP T3. Material evaluation3.1 

Q13. What is the adaptable technique for predicting shipping?

For logistic and transportation planning, Lee (2016) uses ARM to extract purchase patterns and perform if-then-else rules to predict customerAC CEPT EDM ANUS CRIP Tpurchasing behaviour, thus proposing the GA approach to optimize anticipatory shipping assignment.