M.S. Applied Data Science - Capstone Chronicles 2025

4

opportunity to employ contemporary machine learning methods to uncover patterns in cross shopping behavior. Problem Identification and Motivation This project focuses on addressing the challenges of local business engagement in San Diego County. With the highest cost of living in the United States, it is important for local and nationwide businesses in the county to find meaningful ways to connect with their community and neighboring businesses. By analyzing San Diego as a network of brands and shared customers, businesses can quickly identify local brand collaboration and improve both business and community impact. Using network analysis, we seek to identify latent brand partnership opportunities for San Diego-based businesses by leveraging machine learning methods to uncover shared customer bases and detect competitive clusters. By uncovering these connections, this analysis seeks to support businesses by providing alignment with established brands and enabling mutual growth. Over time, model results could also be used to inform local policy, such as requiring support partnerships along with new, large retail development permits, promoting sustainable growth in the community. Definition of Objectives Our goal is to build a directed graph to model the relationship between local brands based on their overlapping customers. This will allow for the discovery of the clusters (i.e., brand ecosystems) that exist in the region and reveal the market segmentation structure in the area. Additionally, we will identify the brands that appear to be “magnets” or “central hubs” in their business category. Lastly, for any given business, we propose a methodology to identify which brands are their strongest first-degree and second-degree connections in their given locale. This

212