This repository features advanced analytics and AI projects applied to the financial sector and industrial process optimization.
- Context: Critical quality variables in the plant relied on laboratory tests with response times of several hours.
- Solution: Designed a virtual sensor (Soft Sensor) using regression models to estimate quality in real-time based on process variables (Pressure, Temperature, Speed).
- Impact: Waste reduction and production line optimization by enabling real-time process adjustments.
⚠️ Confidentiality Note: The source code and datasets for this project are the private property of Proquinal S.A. Only methodology and impact are documented here under strict data handling ethical standards.- Tech Stack: Python, Scikit-Learn, Regression, Process Engineering.
- Context: High volatility and the vast number of assets in the stock market make it difficult to create balanced investment portfolios.
- Solution: Implemented dimensionality reduction (PCA) and clustering (K-Means) techniques to group S&P 500 stocks based on their risk-return behavior.
- Impact: Identified 4 distinct investment profiles, enabling strategic diversification based on data rather than assumptions.
- Tech Stack: Python, Pandas, Scikit-learn, Matplotlib.
- Context: Financial telemarketing campaigns often suffer from low conversion rates, wasting resources on low-probability leads.
- Solution: Developed a classification model based on Random Forest to predict if a client will subscribe to a term deposit.
- Impact: The model allows for lead prioritization, increasing campaign efficiency and optimizing analyst time.
- Tech Stack: Python, Random Forest, Seaborn.