Portuguese Bank Direct Marketing Optimization
Customer Lead Scoring & Campaign ROI Maximization
Key contributions
- Mitigated severe class imbalance between subscribers and non-subscribers.
- Identified call duration, prior campaign status, and Euribor index as paramount conversion signals.
- Formulated decile-based scoring framework prioritizing high-propensity outreach.
Applied Machine Learning Engineering Specification
Structured technical analysis conforming to production machine learning standards.
Problem Formulation
Direct telemarketing campaigns suffered from low conversion rates and escalating operational costs caused by untargeted cold calling across the entire customer base.
Data Ingestion & Feature Space
Comprehensive Portuguese banking dataset comprising 45,211 direct marketing interactions with 17 attributes across client demographics, balance levels, and economic indices.
Preprocessing & Methodology
Conducted exhaustive exploratory data analysis, handled non-linear feature interactions, applied categorical encoding, and addressed class imbalance.
Model Architecture & Training
Trained and benchmarked Logistic Regression, Decision Trees, Random Forests, and Gradient Boosted Classifiers with hyperparameter optimization.
Validation & Evaluation Protocol
Prioritized Precision-Recall AUC and F1-score over raw accuracy to prevent false-negative lead loss, utilizing confusion matrices and threshold tuning.
Empirical Results & Discoveries
Isolated the top 20% lead tier capable of generating over 65% of successful deposit subscriptions, identifying economic sentiment as a key driver.
Production & Business Impact
Supplies marketing operations with an empirical lead-scoring system that substantially cuts campaign costs while raising net conversion yield.
Tech stack
Want to see the source code?
Access the complete Jupyter notebook, preprocessing routines, and evaluation logs.