Air_Quality_Analysis

🌫️ Air Quality Prediction using RNN

A data-driven website to forecast PM1, PM2.5, and PM10 levels across Indian cities using meteorological data from IMD, CPCB, and SNDTW UMIT.
It implements Linear Regression and Recurrent Neural Networks (RNN) to deliver:

• 📈 Real-time AQI forecasting
• 🩺 Health recommendations
• 🤖 Chatbot to answer queries on pollution & risks
• 🌍 Scalable design for multi-city deployment

Achieved 76%+ accuracy in both Mumbai and Cuttack using RNN models.

---

🎥 Demo

📁 Google Drive Backup:
Click to open Drive folder

---

📊 Results & Observations

🏙️ Mumbai

PM2.5 Equation:

PM2.5 = 0.0000 * INDEX + 3.3150 * MAX - 4.6039 * MIN - 2.5933 * AW + 48.1042

PM10 Equation:

PM10 = 0.0000 * INDEX + 7.6420 * MAX - 6.9687 * MIN - 4.1669 * AW + 15.7419

📈 R² Score: 0.7611

✔️ Positive impact of MAX temperature
❌ Negative impact of MIN temperature and wind speed (AW)

---

🌾 Odisha - Cuttack

PM2.5 Equation:

PM2.5 = 0.0000 * INDEX - 0.1707 * MAX - 3.0020 * MIN + 97.5438

PM10 Equation:

PM10 = 0.0000 * INDEX + 0.2133 * MAX - 10.1975 * MIN + 295.3829

📈 R² Score: 0.7569

❌ Strong negative correlation with MIN temperature
➕ Slight positive correlation with MAX
🚫 INDEX was found statistically insignificant

---

⚙️ Features

• 🔍 AQI predictions for PM1, PM2.5, PM10
• 👥 Personalized health-based feedback
• 🤖 Chatbot for environmental awareness
• 📊 Linear Regression & RNN models
• 📱 Clean UI & mobile-friendly experience

---