A Raspberry Pi Based Real-Time Weather Monitoring and Short-Term Prediction System using IoT, InfluxDB, Open-Meteo and Machine Learning

A complete end-to-end IoT + Cloud + Machine Learning project that collects real-time weather data, stores it in a time-series database, performs short-term predictions, and visualizes everything on a live web dashboard.

---

Project Overview

This project demonstrates how modern IoT systems integrate edge devices, cloud databases, REST APIs, machine learning models, and web dashboards into a single, scalable architecture.

The system:

• Collects real-time weather data from Open-Meteo API
• Runs a continuous ingestion service on Raspberry Pi
• Stores data in InfluxDB Cloud (time-series database)
• Exposes data via FastAPI backend
• Predicts temperature & humidity for the next 30 minutes using ML
• Visualizes live + predicted data on a Netlify-hosted dashboard

---

Key Features

• Real-time weather monitoring
• Raspberry Pi based data ingestion
• Cloud-native time-series storage (InfluxDB)
• FastAPI powered REST backend
• Machine Learning based short-term prediction
• Interactive live dashboard (Chart.js)
• Deployed using modern DevOps tools

---

System Architecture

Open-Meteo API
      ↓
Raspberry Pi (Ingestion Service)
      ↓
InfluxDB Cloud (Time-Series Storage)
      ↓
FastAPI Backend (REST API + ML)
      ↓
Netlify Frontend (Dashboard)

Live Project Demo

You can visit the live deployed dashboard here:

Live Dashboard:
https://raspberrypi-weather-dashboard.netlify.app/

The dashboard displays:

• Real-time weather data
• Interactive charts and trends
• Short-term temperature & humidity predictions
• AI-powered forecast indicators

Note: Data updates automatically based on the Raspberry Pi ingestion service and backend API.

Project Screenshots

User Authentication – Login

User Authentication – Sign Up

Weather Analytics Dashboard (Live View)

Real-Time Temperature Trend & Forecast

InfluxDB Cloud – Time-Series Data Storage

Data Ingestion Layer (Raspberry Pi)

• Runs continuously on Raspberry Pi
• Fetches live weather data from Open-Meteo API
• Pushes data every 5 minutes to InfluxDB

---

Project Structure

raspberrypi-weather-iot-ml/
│
├── backend/        # FastAPI + ML backend
│   ├── app/
│   │   ├── main.py
│   │   ├── api/
│   │   │   └── weather.py
│   │   ├── core/
│   │   │   └── config.py
│   │   └── services/
│   │       └── influx_service.py
│   └── ml/
│       ├── data_loader.py
│       ├── preprocess.py
│       ├── predict.py
│       ├── train_baseline.py
│       ├── temp_model.pkl
│       └── hum_model.pkl
│
├── ingestion/      # Raspberry Pi data ingestion
│   └── open_meteo.py
│
├── database/       # InfluxDB helpers
│   └── influx.py
│
├── frontend/       # Web dashboard
│   ├── index.html
│   ├── style.css
│   └── script.js
│
├── config/         # Environment configuration
├── .gitignore
├── README.md
└── LICENSE

---

Key File:

ingestion/open_meteo.py

---

Database Layer (InfluxDB Cloud)

• Time-series optimized storage
• Stores temperature, humidity, pressure, wind speed, wind direction
• Supports high-frequency writes & queries

Measurement Example:

• weather_live

---

⚙️ Backend Layer (FastAPI)

The backend serves as the central brain of the system.

Available API Endpoints

Endpoint	Description
/	Health check
/api/weather/latest	Latest live weather data
/api/weather/predict	ML-based 30-minute prediction

---

Machine Learning Module

Objective

Predict temperature and humidity for the next 30 minutes.

Approach

• Model: Linear Regression
• Input: Last 60 minutes (12 data points)
• Output: Temperature & Humidity (t + 30 min)

ML Pipeline

InfluxDB → Data Loader → Preprocessing → Model → Prediction

---

Frontend Dashboard

• Built using HTML, CSS, JavaScript
• Uses Chart.js for visualization
• Auto-refreshes live data every 5 seconds
• Updates predictions every 30 seconds

Deployment: Netlify

---

Deployment Summary

Component	Platform
Ingestion	Raspberry Pi
Database	InfluxDB Cloud
Backend	Render (API)
Frontend	Netlify

---

Environment Variables

All sensitive credentials are stored using .env:

INFLUX_URL=...
INFLUX_TOKEN=...
INFLUX_ORG=...
INFLUX_BUCKET=...

---

Testing & Validation

• Verified live data flow from Raspberry Pi to InfluxDB
• Validated API responses via browser & fetch calls
• Compared dashboard values with InfluxDB Data Explorer

---

Academic & Practical Relevance

• IoT Architecture Design
• Cloud Databases (Time-Series)
• RESTful API Development
• Machine Learning Integration
• Full-Stack Deployment

---

Future Improvements

• Replace Linear Regression with LSTM/GRU
• Add anomaly detection
• Deploy backend on Raspberry Pi locally
• Add alerting system (SMS / Email)
• Secure APIs with authentication

---

Author

Shahriar Alom Masud
B.Sc. Engg. in IoT & Robotics Engineering
University of Frontier Technology, Bangladesh
Email: shahriar0002@std.uftb.ac.bd
LinkedIn: https://www.linkedin.com/in/shahriar-alom-masud

---

License

This project is licensed under the MIT License.

---

If you like this project, give it a star on GitHub!