Year 2021, Issue 143

Date published

15.7.2021

Table of content

  • Todor Krystevich
    MARKET BASKET ANALYSIS USING R
    JEL: M3, C8, C40.
    Keywords: market basket analysis, sales transactions, association rule mining, descriptive modelling, cross-selling.
    Summary: ction 9 1. Concept and Tools 13 1.1. Defining the problem 13 1.2. Conceptual framework 15 1.3. Basic concepts 17 1.4. Logical data model 18 1.5. Exploratory analysis of sales transactional data 21 1.6. Transaction-class objects 25 1.7. Descriptive analysis of sales transactional data 27 1.8. Market basket analys ... ... Introduction 9 1. Concept and Tools 13 1.1. Defining the problem 13 1.2. Conceptual framework 15 1.3. Basic concepts 17 1.4. Logical data model 18 1.5. Exploratory analysis of sales transactional data 21 1.6. Transaction-class objects 25 1.7. Descriptive analysis of sales transactional data 27 1.8. Market basket analysis metrics 32 1.9. Association rule mining 40 1.9.1. Identifying common subsets 41 1.9.2. Identifying strong association rules 47 1.10. Visual analysis of market basket 51 1.10.1. Visualization of absolute and relative frequency 53 1.10.2. Visualization of market basket analysis metrics 54 1.10.3. Visualization of association rules 60 2. Big Data Market Basket Analysis 69 2.1. Analytical approach 69 2.2. Data 72 2.2.1. Import and conversion 72 2.2.2. Overview of transactions, assortment items and product groups 73 2.3. Big Data exploratory data analysis 74 2.4. Generating rules and making recommendations 81 2.5. Visual analysis and interpretation 89 2.6. Extended market basket analysis taking contribution margin into account 98 Conclusion 105 References 109 The digital transformation of business, the emerging customer-centric business models as well as advanced tools and techniques for analysis imply the need for a marketing culture focused on data and analysis. This monograph explores and presents approaches, algorithms, techniques and procedures for analyzing the market basket with data from sales transactions. Although the statement is based on solid theoretical sources, the main emphasis is placed on the methodology and extraction of patterns from large datasets by using association rules. Guided by the understanding of the leading role of transactional data from sales transactions in mining personalized information and patterns of customer behaviour, our goal is to provide the reader with specific practical guidelines and reproducible operating procedures for market basket analysis with the R programming language. The overall approach of the study is based on specific examples, work instructions, program code and guidelines for interpreting the results in the context of supporting marketing decisions in commercial enterprises.