An interactive dashboard for analyzing and visualizing wildfire occurrences across Canada (2013–2023). Features dynamic filtering by province and year range, with various data visualizations including maps, charts, and summary statistics.

Python package for calculating Air Quality Index (AQI) based on CONAMA Resolution 491/2018 standards. Supports PM10, PM25, SO2, NO2, CO, O3.

FETCH (Framework for Environmental Type Classification Hub) is a QGIS-based tool for automated Local Climate Zone (LCZ) classification. It combines Google Solar API data acquisition with advanced geospatial processing to analyze urban morphology and climate characteristics.

A suite of simplified climate analysis and urban comfort tools built on Ladybug Tools, providing an intuitive interface for complex environmental studies.

High-precision ML regression model predicting vehicle CO2 emissions from technical specifications. Achieves R²=0.9972 with Random Forest using Transport Canada data.

Interactive Tableau dashboard exploring global CO₂ emissions trends, rankings, and growth from 2000–2022 using public OWID climate data.

A comprehensive web application for environmental site analysis providing vegetation, terrain, and climate data visualization. Built with Next.js and Django.

Projeto dedicado a investigar os incêndios florestais no Brasil durante o ano de 2024. Através da análise de dados climáticos e informações sobre o risco de fogo, o projeto busca compreender como variáveis como dias sem chuva e precipitação influenciam a ocorrência de incêndios.

A machine learning system that predicts Water Quality Index (WQI) and Water Quality Classification (WQC) using engineered water-quality features. The project includes full preprocessing with missing-value handling, outlier treatment, dynamic WQI calculation, and both regression and classification modeling powered by XGBoost and Scikit-learn.

A machine learning project that predicts car CO2 emissions based on engine size using linear regression. Features data visualization, model training, and performance evaluation with scikit-learn and matplotlib.

Previous works on smart surveillance systems often only focused on one task such as violence detection, and were heavyweight systems at the same time. This project aims to create a smart surveillacne system that does more than an average human (or a team of humans) could ever do.