Data-driven E-commerce UI Personalization: Going Beyond Product Recommendations
Description
The dataset includes 1. online store customer behavior data (clickstream) from 1.04.-30.11.2023, used to cluster customers and evaluate the effectiveness of implemented modifications (catalog: learning-dataset) 2. clustering results to verify the effectiveness of implemented changes (catalog: clustering) 3. detailed data for calculation of macro-conversion indicators (catalog: macro-conversion-indicators) 3. detailed data for calculation of micro-conversion indicators (catalog: micro-conversion-indicators)
Files
Steps to reproduce
Behavioral data is the basis for clustering. The study used the K-means algorithm, with k=4. The results of macro-conversion and micro-conversion are calculated based on the corresponding customer actions included in the learning dataset. Actions included in the calculation of PCRs are derived from the expected customer journey.