WeatherNet

Important Links

Module	Link
Website	WeatherNet-il
Video	Video
Github	WeatherNet Project
User Manual	User Manual
Developer Manual	Developer Manual
Phase A Paper	Phase A Paper
Phase B Paper	Phase B Paper

About WeatherNet

WeatherNet is an advanced weather forecasting system designed to provide accurate mid-term temperature predictions for Israel. It leverages machine learning to analyze historical and real-time data, delivering reliable forecasts through a user-friendly web interface.

Project Phases

The development of WeatherNet was divided into two primary phases:

• Phase A (Research & Proof of Concept - POC):
  Conducted in-depth research on weather forecasting methods and machine learning techniques, culminating in a proof-of-concept model to validate feasibility.

• Phase B (Development & Implementation):
  Transitioned from research to full-scale implementation, including training the ML model, developing the user interface, and deploying the system.

System Components

WeatherNet consists of two primary components:

• Backend Machine Learning Model:
  Processes weather data, trains on historical and real-time information, and generates accurate predictions.

• Frontend Web-Based Platform:
  Provides an interactive interface for users to access forecasts, compare results, and explore model insights.

Machine Learning Architecture

Our model follows a hybrid approach combining:

• 1D Convolutional Neural Networks (CNNs) for feature extraction.
• Positional Encodings for both spatial and temporal context.
• Transformer Encoder for modeling complex relationships between stations and across time.
• Fully Connected Layers for generating final predictions.

This architecture enables high-accuracy weather forecasts by capturing both temporal dependencies and geographical relationships.

Team Members

• Yuval Rozner
• Dor Shabat

Tools and Technologies

• Machine Learning & Data Processing: Python, PyTorch, NumPy, Pandas, Scikit-learn
• Frontend Development: React, Material-UI, Styled-Components
• Backend & Deployment: Firebase Hosting, Firebase Functions, Git

Contact

• Yuval: yuvalrozner98@gmail.com
• Dor: dorshabat55@gmail.com