Applying the Smart Literature Review to supply chain and logistics research

Abstract

Purpose

We apply a smart literature review to supply chain and logistics research to demonstrate approaches to the use of machines in academic research projects.

Research Approach

The work uses machine learning and natural language processing in python notebooks to illustrate data pipelines for search breadth across multiple databases. Code and human checks are identified and tested throughout the stages of the smart literature review such as data cleaning, testing for false positives, unsupervised machine learning for topic identification and summarising topic areas.

Findings and Originality

We present a reproducible smart literature of supply chain and logistics research. The work successfully shows how to overcome many challenges in data standardisation across literature databases and in using unsupervised machine models of topics in supply chain and logistics research. Finally, we present conclusions on key aspects of topic modelling and using NLP in supply chain research.

Research Impact

Our model contributes significantly to the smart literature review methodology, with specific application to supply chain and logistics research, fostering a new understanding of AI's capacity to synthesise and interpret extensive academic literature.

Practical Impact

AI is changing the balance of resource use in research projects and this work shows ways in which the smart literature review will become a basic element of future projects.