PORTFOLIO

Developed a predictive model to identify bank customers most likely to accept a personal loan offer. By targeting the right customers, Citizens Bank can optimize marketing spend, improve conversion rates, and reduce wasted outreach.

• Citizens Bank wanted to improve the efficiency of personal loan marketing campaigns.
• Traditional blanket marketing was costly and resulted in low conversion rates.
• The goal was to predict which customers were most likely to accept a personal loan offer so that future campaigns could be more targeted and precise.

• Cleaned and preprocessed customer records from a dataset of 5,000 customers. This included addressing data entry errors and irrelevant features.
• Performed in-depth Exploratory Data Analysis (EDA) to explore distributions and understand the key drivers of loan acceptance.

• Implemented a Decision Tree Classifier, a transparent and interpretable model well-suited for this classification task.
• Utilized a standard train-test split to prepare the data for training and evaluation.
• Pruned the Decision Tree to prevent overfitting, enhancing the model's ability to generalize to new, unseen customer data.

• Generated a confusion matrix and calculated key classification metrics—including accuracy, precision, and recall—to thoroughly assess model performance.
• Analyzed customer attributes to identify the characteristics most associated with a higher likelihood of accepting a loan, such as income, education level, and account ownership.

• Higher Conversion Rates: Targeting only high-likelihood customers can significantly boost marketing ROI.
• Lower Acquisition Costs: Fewer wasted outreach efforts.
• Scalability: Model can be retrained with updated data to adapt to changing market conditions.

• Integrate into the bank’s CRM for real-time targeting.
• Expand features with transaction-level data for even greater predictive power.
• Conduct A/B testing to quantify uplift from targeted campaigns.