Identification of Customer Behavioral Attributes through Topic Modelling

Explore a 35-minute conference talk from the Data Science Festival on identifying customer behavioral attributes through topic modeling. Discover how dunnhumby leverages advanced techniques to analyze retail basket data and gain insights into shopping behavior. Learn how the company reimagines baskets as textual documents and applies topic modeling to multiple baskets for identifiable customers. Understand the process of training a topic model on millions of customers using transaction data from a North American retailer. Gain insights into how the discovered topics correspond to distinct behavioral patterns and can be used to tag customers with descriptive shopping behaviors. Delve into this innovative approach that shifts the definition of a document to incorporate multiple baskets, providing a deeper understanding of individualized shopping habits.