Big Data Analytics of Supply Chains with Process Mining

Abstract

The purpose of this paper is to apply process mining techniques on a supply chains dataset obtained from data.mendeley.com. This dataset is previously used by the company DataCo Global [1]. However, Machine Learning and R programming language were used in [1]. What distinguishes our work from [1] is using Fluxicon Disco as process mining platform to further investigate and analyze the behaviors, shopping trends and demands of customers from 164 different countries. The dataset contains variables such as products, category types, payment type, delivery status, order status, shipping mode, etc. The dataset was collected within a period of three years (2015 - 2018). The dataset includes: 164 cases from worldwide customers who ordered products [online] from a website. In total, there are 180,519 events or the number of times when customers have ordered something from the website, and there are 118 activities or different ways to handle a customer order. Due to big size of the dataset used in this research, the study can be somehow relevant to big data analytics as well. The results of the study revealed interesting information about the demographic types of the customers [globally] in addition to their most preferred purchased items, payments methods, delivery status and so on.