Featured Work

XANE Chatbot

An AI-powered assistant for interacting with my portfolio, answering queries, and processing files.

Rule-based system with LLM integration via OpenRouter API
Handles text, images, and PDFs using Supabase for storage
Image generation using Flux and Turbo via Pollinations
Built with Streamlit for a seamless user experience

Python Streamlit Supabase OpenRouter

Semi-Automated Statistical Modeling Toolkit

A comprehensive suite of statistical modeling tools focusing on rigorous analysis, proper diagnostics, and interpretable results rather than pure predictive performance.

Three specialized modules for Linear Regression, Logistic Regression, and Time Series Analysis
Emphasis on statistical assumptions, diagnostic checks, and model interpretation
Automated model diagnostics including multicollinearity, normality, and heteroscedasticity checks
Interactive visualizations for model evaluation and assumption verification
Detailed explanations of statistical concepts and results interpretation

Python Statsmodels Streamlit Plotly

Fixed Effects Panel Data Analysis of Spanish Dairy Farms

Econometric analysis of dairy farm productivity using panel data techniques.

Analyzed panel data (247 farms, 1,482 observations) using Cobb-Douglas and Translog models in R
Identified key drivers (Cows: 0.66%, Feed: 0.38%, p < 0.001) with increasing returns to scale (RTS = 1.11)
Ensured robustness with clustered errors and log transformation, achieving ~80% adjusted R²

View on GitHub

View Research Paper

Time Series Analysis of the Volatility Index

Forecasting and modeling of market volatility using advanced time series techniques.

Applied AR(I) and cubic models to daily VIX data (1990–2025) in R, with 6.67% MAPE
Confirmed stationarity (p = 0.188) and validated with Box-Ljung and Box-Pierce tests
Forecasted VIX levels (19.3–28.65) through 2034 with confidence intervals

View on GitHub

View Research Paper