Weather forecasting with Neural Networks

Weather Forecasting with Neural Networks is a machine learning project focused on predicting climate variables in León de los Aldama, Mexico, using historical meteorological data collected from seven weather stations across the city. The project explores the application of dense neural networks to analyze historical climate patterns and forecast temperature and precipitation trends for 2025.

Key Highlights

  • Meteorological Data Analysis: Collected and analyzed historical records from seven weather stations across León, including maximum temperature, minimum temperature, precipitation, and evaporation.
  • Neural Network Modeling: Designed a three-layer dense neural network to learn patterns from historical weather measurements.
  • Climate Forecasting: Explored the prediction of key meteorological variables over a 12-month forecasting horizon.
  • Data Preprocessing: Organized and prepared historical measurements from stations with different recording periods and geographical characteristics.
  • Scientific Research: Documented the methodology and project findings in a technical report written in LaTeX.

Tech Stack & Methodology

  • Machine Learning (TensorFlow + Keras): Developing neural network models for analyzing historical meteorological data and generating climate predictions.
  • Data Processing (Python + Pandas): Cleaning, organizing, and manipulating weather records to prepare datasets for model training.
  • Model Development (Jupyter Notebook): Conducting experiments, exploring data, and evaluating the neural network development workflow.
  • Scientific Documentation (LaTeX): Producing a technical research report describing the project's objectives, data sources, and methodology.
something cherry