propensity-modeling

End-to-end e-commerce product analytics platform with BigQuery SQL, Tableau dashboards, purchase propensity modeling, A/B testing, and an AI analytics copilot.

Propensity model training with XGBoost

ML-powered channel optimization engine for pharma sales reps. Engagement-based NBA with XGBoost propensity models, Streamlit dashboard, and synthetic data for the German pharma market.

Propensity-модель для приоритизации клиентов банковской кампании на реальных данных UCI.

Behavioral embeddings improve energy-program propensity prediction and ranking; propensity neighborhoods add modest calibration benefits.

Lead propensity modelling that ends in a decision: calibrated probabilities, an expected-value threshold derived from the economics, and a generated model card.

Customer segmentation and 90-day repeat-purchase propensity modeling using the UCI Online Retail II dataset.

End-to-end bank marketing analytics and customer propensity modelling using Python, scikit-learn, Random Forest and Power BI to improve campaign targeting and customer prioritization.

Azure Product Growth Intelligence Platform — commercial customer analytics with dbt, SQL, churn, propensity, LTV, attribution, experimentation and Power BI-ready reporting.

A self-hosted BigQuery ML pipeline that predicts purchase propensity from GA4 events and pushes the result back to GA4 as a user property via the Measurement Protocol. Built for Google Ads remarketing. Consent-aware, cost-capped, and production-hardened.

Unit economics on two public datasets: seller-acquisition funnel, customer cohort LTV, freight-driven segment margins, and an out-of-time cross-sell propensity model.

1st-place ODSC AI x Zerve Datathon solution using leakage-safe feature engineering, nine-stage funnel analytics and LightGBM upgrade prediction.

Propensity scoring model for user conversion prediction

Predicts which dormant B2B clients will reactivate if contacted. End-to-end ML pipeline turns 2M transactions into a campaign-ready priority list via Extra Trees + RFM feature engineering. Delivers 3.4× lift over random outreach. Python, scikit-learn, XGBoost.

Machine learning project to predict customer purchase propensity using e-commerce behavioral data

This project segments Starbucks customers using transaction and offer data. Through preprocessing, feature engineering, and clustering (K-Means), it identifies distinct customer groups, providing insights to personalize marketing, improve engagement, and boost customer retention.

Customer propensity scoring, K-means segmentation, and CLV estimation with survival analysis