Data Science in Business: Understanding Growth from a Data-Driven Perspective: A Case Study of a Lisbon-Based Bakery

Dissertation presented as the partial requirement for obtaining a Master's degree in Data Driven Marketing, specialization in Marketing Research and CRM ; In the highly competitive food industry, data analytics has become an essential tool for driving strategic growth and expansion. Leveraging data insights allows businesses to make informed decisions, optimize operations, and enhance customer experiences, thereby building a strong foundation for business growth. This work analyzes customer and sales data from a renowned bakery in Lisbon, Portugal, which contains data from 16 points of sale across four years. The primary objective of this research is to leverage advanced analytics to identify underlying patterns in the data that help the brand to grow. In order to achieve this, four main categories were identified that build the foundation of growth: profit generation, cost reduction, risk mitigation, and improving innovation life cycles. The insights of the analysis aim to facilitate and optimize strategic business decisions. The underlying patterns were identified by using association rule mining, utilizing the apriori algorithm, with which every shop and every year of that bakery was analyzed. Subsequently, the identified rules were further explored through a Meta-Analysis, utilizing K-means clustering to investigate similarities across points of sale based on their association rules. Additionally, Kruskal-Wallis tests were employed to assess the significance of seasonal variations, followed by Dunn’s post-hoc test for a more in-depth analysis. The clusters were further analyzed by examining the coefficient of variation within each cluster to gain deeper insights. The analysis revealed various association patterns that were consistently found across different shops and years. Significance could be found in sales behaviour across the seasons, weekdays, and months. Significant seasonal variations were also observed for the metrics of the association rules. The clusters formed based on the association rules did ...