📊 Analyze customer churn with an end-to-end machine learning system featuring a Streamlit dashboard, model comparison, and explainable AI.

🔍 Streamline tabular binary classification with model interpretability and SHAP consistency analysis for clear insights and robust evaluation.

📊 Analyze customer retention and churn with SQL and Power BI to boost growth through insights on spending patterns and repeat behavior.

💡 Analyze e-commerce data to uncover revenue trends and segment customers for actionable insights, driving strategic business growth.

📊 Predict fraudulent transactions using SQL and Python with labeled data for accurate supervised learning and robust model evaluation.

🎬 Analyze movie reviews using LSTM for accurate sentiment classification, helping you understand viewer opinions with a balanced dataset.

🔍 Analyze IMDB movie reviews using LSTM for binary sentiment classification with high accuracy and custom prediction capabilities.

Built a Logistic Regression model in R to predict housing loan approvals using financial and demographic data. Performed data preprocessing, feature engineering, and model evaluation using ROC curve and confusion matrix. Achieved ~74% accuracy, showcasing the effectiveness of ML in loan decision-making.

Machine learning project for classifying California housing prices into high or low categories using Logistic Regression, Decision Trees, and ensemble methods, with performance evaluation via ROC, AUC, and cross-validation.

Building a lead scoring model using logistic regression to help an education company identify the most promising leads and achieve an ~80% conversion rate.

Customer churn prediction model (Logistic Regression) in Python, scoring and retention action CLI built in Go. / Modelo de prediccion de fuga de clientes con Python y CLI de retencion en Go.

End-to-end SaaS churn risk modeling and revenue exposure analytics pipeline (PostgreSQL + Python + Tableau).

Tool demonstrating building credit risk models

Default-Risk Prediction & Screening at Loan Origination in P2P Consumer Lending, with a Double Machine Learning Extension of the Effects of Longer Terms and High Interest Rates

A single ROC curve for multiclass tasks. Suited for highly imbalanced classes. Below you can find a link to the published pre-print paper with medical and financial applications.

Machine learning classification in Python (logistic regression) for benign vs malignant breast tumor prediction using digitized image-derived features and standard evaluation metrics.

This machine learning project predicts NBA draft prospect success using physical attributes and athletic measurements from the NBA Draft Combine (2000-2025), comparing Random Forest and XGBoost classification models.

ROC curve comparison and hyperparameter tuning for Decision Tree classifiers on two OpenML datasets

Evaluation of Binary Classifiers

A comprehensive machine learning project for spam email detection using multiple classification algorithms and feature representations. This project compares traditional Bag-of-Words (BoW) with modern Sentence Embeddings (SBERT) approaches, achieving up to 96.85% F1-score and 99.93% AUC.