Exploratory Data Analysis

Exploratory data analysis on a UK grocery e-commerce dataset stored in AWS S3 as Parquet files. Credentials managed through environment variables. The dataset spans five tables joined at analysis time.

• arrow_forwardOrders
• arrow_forwardUsers
• arrow_forwardInventory
• arrow_forwardRecurring customers
• arrow_forwardAbandoned carts

• arrow_forwardCustomer behaviour and segmentation
• arrow_forwardUK NUTS1 regional trends
• arrow_forwardProduct performance and pricing patterns
• arrow_forwardUser cohort retention dynamics
• arrow_forwardAverage order value evolution over time
• arrow_forwardCart abandonment patterns

• arrow_forward70% of users made only one purchase, the business is dominated by one-time buyers
• arrow_forward69% of users abandoned at least one cart during their lifetime
• arrow_forwardAOV plateaued at roughly 56 EUR from April 2021 onward
• arrow_forwardCohort retention stabilized around 10-15%, but newer cohorts retained worse than earlier ones
• arrow_forwardCompressing LTV (cumsum of retention x avg orders x AOV) erodes the margin against CAC; a structural challenge where acquisition costs become progressively harder to justify