A selection of data scienceprojects I've built.
Developed a rigorous statistical framework for estimating marketing ROI in insurance and banking using observational data. The methodology combines advanced causal inference techniques with machine learning to accurately measure advertising effectiveness across multiple channels.
Developed an innovative computer vision pipeline that automates the extraction of handwritten data from life insurance applications. The solution combines template matching, region extraction, and CNNs to transform weeks of manual review into minutes, while enabling the use of previously untapped predictive features in underwriting models.
Developed a comprehensive ML solution to predict customer churn for a telecommunications company. The model achieved 92% accuracy and helped reduce churn rate by 18% through targeted retention campaigns.
Created an NLP tool that analyzes sentiment across Arabic, French, and English. The tool processes social media data and customer reviews to provide real-time sentiment insights for businesses operating in diverse markets.
Developed a deep learning system for automated pneumonia detection in chest X-rays. The model achieved 95% accuracy and assists healthcare professionals in diagnosis, potentially improving patient outcomes in resource-limited settings.
Built an intelligent trading assistant combining technical analysis with ML to predict market movements. The platform analyzes financial indicators and news sentiment to provide trading recommendations with risk assessment.
Developed a comprehensive toolkit of reusable Python modules for common data science tasks. This open-source project includes automated EDA, feature engineering pipelines, and model evaluation utilities used by fellow data scientists in the community.
More projects available on GitHub
← Back to home