/Online_Retail_Customer-Segementation

Introduction In ecommerce companies like online retails, customer segmentation is necessary in order to understand customers behaviors. It leverages aqcuired customer data like the one we have in our case, transactions data in order to divide customers into groups. Our goal in this Notebook is to cluster our customers to get insights in: Increasing revenue (Knowing customers who present most of our revenue) Increasing customer retention Discovering Trends and patterns Defining customers at risk We will do RFM Analysis as a first step and then combine RFM with predictive algorithms (k-means). RFM Analysis answers these questions: Who are our best customers? Who has the potential to be converted in more profitable customers? Which customers we must retain? Which group of customers is most likely to respond to our current campaign? More about RFM here.

Primary LanguageJupyter Notebook

online-retail-case

  • Download the dataset Online Retail and put it in the same directory as the iPython Notebooks.
  • EDA notebook which is an exploration of the data.
  • Market Basket Analysis to study customers purchases (Product association rules - Apriori Algorithm).
  • Customer Segmentation to help us divide them into groups. (RFM Analysis - Clustering using K-means)

Perspectives:

  • Classification of new customers into discovered segments.
  • Clustering of transaction dataset based on its initial features (CustomersID, InvoiceDate,etc), apply PCA, feature selection.
  • Create new features (Time, Day of week, Month) to explore customers behavior per time/day.