My Projects

Objective: To develop and evaluate a machine learning model to predict the outcome of NBA games based on historical team data.

Technologies: Python, Pandas for data analysis and cleaning, Scikit-learn for creating and training models (e.g., Logistic Regression, Random Forest).

Challenges & Results: The main challenge was feature engineering to identify the most influential statistics. The final model achieved a predictive accuracy of 71%, demonstrating the effectiveness of an analytical approach in sports.

Objective: To compare the performance of DuckDB with its spatial extension against PostGIS, analyzing the execution times of complex geospatial queries on OpenStreetMap data.

Technologies: Python, DuckDB, GeoPandas, PostgreSQL/PostGIS, Jupyter Notebook for results analysis.

Challenges & Results: Setting up a consistent testing environment was crucial. The results highlighted DuckDB's remarkable speed for in-process analytical workloads, providing useful data for choosing a DB in future projects.

Objective: To create an interactive and responsive web application to test users' knowledge on various topics, with immediate feedback on their answers.

Technologies: HTML5, CSS3, JavaScript (ES6). Special attention was given to DOM manipulation for a dynamic user experience.

Challenges & Results: Managing the application's state (score, current question) in pure JavaScript was a formative challenge. The result is a lightweight, functional, and easily extendable app with new questions.