Build the features to cluster

A: compute Recency, Frequency, Monetary per customer from invoices. B: build one 24-value load profile per day. Concepts: Feature construction for segmentation; aggregation to the right grain Project A: Online Retail II: ~5k customers after cleaning; log1p + standardise R, F, M. Project B: ~1,400 days × 24 hours from the EDA dataset; normalise each day by its mean (shape) and keep daily mean and temperature separately.