app.py

import streamlit as st
import pandas as pd
import numpy as np
import requests
from bs4 import BeautifulSoup
import folium
from streamlit_folium import folium_static
import plotly.express as px
import plotly.graph_objects as go
from datetime import datetime, timedelta
import json
from io import StringIO
import streamlit.components.v1 as components
import base64
from sklearn.ensemble import RandomForestRegressor
from prophet import Prophet
import tensorflow as tf
from xgboost import XGBRegressor
import seaborn as sns
from plotly.subplots import make_subplots

# Define utility functions first
def download_csv(df, filename):
    """Generate a download link for a dataframe"""
    csv = df.to_csv(index=False)
    b64 = base64.b64encode(csv.encode()).decode()
    href = f'<a href="data:file/csv;base64,{b64}" download="{filename}.csv">Download {filename} CSV</a>'
    return href

# Page configuration
st.set_page_config(layout="wide", page_title="Pakistan Climate & Disaster Monitor", page_icon="🌍")

class DataCollector:
    def __init__(self):
        self.cities = {
            'Islamabad': {'lat': 33.7294, 'lon': 73.0931},
            'Karachi': {'lat': 24.8607, 'lon': 67.0011},
            'Lahore': {'lat': 31.5204, 'lon': 74.3587},
            'Peshawar': {'lat': 34.0151, 'lon': 71.5249},
            'Quetta': {'lat': 30.1798, 'lon': 66.9750},
            'Multan': {'lat': 30.1575, 'lon': 71.5249},
            'Faisalabad': {'lat': 31.4504, 'lon': 73.1350},
            'Rawalpindi': {'lat': 33.6007, 'lon': 73.0679},
            'Gwadar': {'lat': 25.1216, 'lon': 62.3254},
            'Hyderabad': {'lat': 25.3960, 'lon': 68.3578}
        }

    def fetch_weather_data(self):
        """Fetch weather data from OpenMeteo"""
        weather_data = []
        for city, coords in self.cities.items():
            try:
                url = f"https://api.open-meteo.com/v1/forecast?latitude={coords['lat']}&longitude={coords['lon']}&hourly=temperature_2m,relativehumidity_2m,precipitation,windspeed_10m&daily=temperature_2m_max,temperature_2m_min,precipitation_sum&timezone=auto&past_days=7"
                response = requests.get(url)
                data = response.json()
                
                # Hourly data
                hourly_df = pd.DataFrame({
                    'datetime': pd.to_datetime(data['hourly']['time']),
                    'temperature': data['hourly']['temperature_2m'],
                    'humidity': data['hourly']['relativehumidity_2m'],
                    'precipitation': data['hourly']['precipitation'],
                    'wind_speed': data['hourly']['windspeed_10m']
                })
                
                # Daily data
                daily_df = pd.DataFrame({
                    'date': pd.to_datetime(data['daily']['time']),
                    'temp_max': data['daily']['temperature_2m_max'],
                    'temp_min': data['daily']['temperature_2m_min'],
                    'precipitation_sum': data['daily']['precipitation_sum']
                })
                
                weather_data.append({
                    'city': city,
                    'hourly': hourly_df,
                    'daily': daily_df,
                    'coords': coords
                })
            except Exception as e:
                st.error(f"Error fetching weather data for {city}: {e}")
                continue
        
        return weather_data if weather_data else None
    
    def fetch_usgs_earthquake_data(self):
        """Fetch earthquake data from USGS website"""
        try:
            # USGS API endpoint for past month's earthquakes
            url = "https://earthquake.usgs.gov/earthquakes/feed/v1.0/summary/2.5_month.geojson"
            response = requests.get(url)
            data = response.json()
            
            # Filter for Pakistan region
            pakistan_data = {
                "type": "FeatureCollection",
                "features": [
                    feature for feature in data["features"]
                    if 60.878 <= feature["geometry"]["coordinates"][0] <= 77.840
                    and 23.692 <= feature["geometry"]["coordinates"][1] <= 37.097
                ]
            }
            return pakistan_data
        except Exception as e:
            st.error(f"Error fetching earthquake data: {e}")
            return None

    def fetch_air_quality_data(self):
        """Fetch air quality data from OpenMeteo"""
        aqi_data = []
        for city, coords in self.cities.items():
            try:
                url = f"https://air-quality-api.open-meteo.com/v1/air-quality?latitude={coords['lat']}&longitude={coords['lon']}&hourly=pm10,pm2_5,carbon_monoxide,nitrogen_dioxide,ozone&timezone=auto&past_days=7"
                response = requests.get(url)
                data = response.json()
                
                df = pd.DataFrame({
                    'datetime': pd.to_datetime(data['hourly']['time']),
                    'PM10': data['hourly']['pm10'],
                    'PM2.5': data['hourly']['pm2_5'],
                    'CO': data['hourly']['carbon_monoxide'],
                    'NO2': data['hourly']['nitrogen_dioxide'],
                    'O3': data['hourly']['ozone'],
                    'city': city
                })
                aqi_data.append(df)
            except Exception as e:
                st.error(f"Error fetching AQI data for {city}: {e}")
                continue
        
        return pd.concat(aqi_data, ignore_index=True) if aqi_data else None

    def create_ml_features(self, weather_data):
        """Create features for ML predictions"""
        features_df = pd.DataFrame()
        for city_data in weather_data:
            df = city_data['hourly'].copy()
            df['city'] = city_data['city']
            
            # Create time-based features
            df['hour'] = df['datetime'].dt.hour
            df['day'] = df['datetime'].dt.day
            df['month'] = df['datetime'].dt.month
            df['day_of_week'] = df['datetime'].dt.dayofweek
            
            # Create lag features
            df['temp_lag_1'] = df['temperature'].shift(1)
            df['temp_lag_24'] = df['temperature'].shift(24)
            
            # Create rolling means
            df['temp_rolling_mean_6h'] = df['temperature'].rolling(window=6).mean()
            df['temp_rolling_mean_24h'] = df['temperature'].rolling(window=24).mean()
            
            features_df = pd.concat([features_df, df])
        
        return features_df.dropna()

    def download_csv(df, filename):
        """Generate a download link for a dataframe"""
        csv = df.to_csv(index=False)
        b64 = base64.b64encode(csv.encode()).decode()
        href = f'<a href="data:file/csv;base64,{b64}" download="{filename}.csv">Download {filename} CSV</a>'
        return href

        
def create_3d_visualization(earthquake_data=None, weather_data=None):
    """Create an interactive 3D visualization of Pakistan"""
    threejs_html = """
    <div id="visualizationContainer" style="width: 100%; height: 600px;">
        <div style="position: absolute; top: 10px; left: 10px; background: rgba(255,255,255,0.8); padding: 10px; border-radius: 5px;">
            <h3 style="margin: 0;">Pakistan Terrain Map</h3>
            <p style="margin: 5px 0;">🔴 Cities | 🟡 Earthquake Events</p>
        </div>
    </div>
    <script src="https://cdnjs.cloudflare.com/ajax/libs/three.js/r128/three.min.js"></script>
    <script src="https://cdnjs.cloudflare.com/ajax/libs/dat-gui/0.7.7/dat.gui.min.js"></script>
    <script>
        const container = document.getElementById('visualizationContainer');
        const scene = new THREE.Scene();
        scene.background = new THREE.Color(0xc6e6ff);  // Light blue sky
        const camera = new THREE.PerspectiveCamera(75, container.clientWidth / container.clientHeight, 0.1, 1000);
        const renderer = new THREE.WebGLRenderer({ antialias: true });
        renderer.setSize(container.clientWidth, container.clientHeight);
        container.appendChild(renderer.domElement);

        // Pakistan outline coordinates (simplified)
        const pakistanOutline = [
            [75.3, 35.5], [72.8, 36.5], [71.2, 34.7], [69.5, 34.0],
            [66.8, 31.2], [63.2, 26.8], [61.5, 25.0], [63.5, 23.5],
            [66.7, 24.5], [67.8, 24.8], [68.9, 27.0], [70.5, 28.5],
            [72.0, 30.0], [74.0, 32.5], [75.3, 35.5]
        ];

        // Create Pakistan terrain
        const shape = new THREE.Shape();
        pakistanOutline.forEach((point, index) => {
            if (index === 0) shape.moveTo(point[0] - 67, point[1] - 30);
            else shape.lineTo(point[0] - 67, point[1] - 30);
        });

        const geometry = new THREE.ExtrudeGeometry(shape, {
            depth: 2,
            bevelEnabled: true,
            bevelSegments: 2,
            steps: 2,
            bevelSize: 0.3,
            bevelThickness: 0.3
        });

        const material = new THREE.MeshPhongMaterial({
            color: 0x4CAF50,
            flatShading: true,
            side: THREE.DoubleSide
        });

        const pakistan = new THREE.Mesh(geometry, material);
        scene.add(pakistan);

        // Add major cities
        const cities = {
            'Islamabad': [73.0931, 33.7294],
            'Karachi': [67.0011, 24.8607],
            'Lahore': [74.3587, 31.5204],
            'Peshawar': [71.5249, 34.0151],
            'Quetta': [66.9750, 30.1798]
        };

        Object.entries(cities).forEach(([name, coords]) => {
            const cityMarker = new THREE.Mesh(
                new THREE.SphereGeometry(0.3, 32, 32),
                new THREE.MeshPhongMaterial({ color: 0xff0000 })
            );
            cityMarker.position.set(coords[0] - 67, coords[1] - 30, 2.5);
            scene.add(cityMarker);

            // Add city label
            const div = document.createElement('div');
            div.className = 'label';
            div.textContent = name;
            div.style.color = 'black';
            div.style.position = 'absolute';
            div.style.padding = '2px';
            div.style.backgroundColor = 'rgba(255,255,255,0.7)';
            div.style.borderRadius = '3px';
            container.appendChild(div);
        });
        """

    if earthquake_data:
        threejs_html += """
        // Add earthquake markers
        const earthquakes = """ + json.dumps(earthquake_data['features']) + """;
        earthquakes.forEach(quake => {
            const coords = quake.geometry.coordinates;
            const magnitude = quake.properties.mag;
            
            const quakeMarker = new THREE.Mesh(
                new THREE.SphereGeometry(magnitude * 0.2, 32, 32),
                new THREE.MeshPhongMaterial({
                    color: 0xFFD700,
                    opacity: 0.8,
                    transparent: true
                })
            );
            quakeMarker.position.set(coords[0] - 67, coords[1] - 30, 2.5);
            scene.add(quakeMarker);
        });
        """

    threejs_html += """
        // Lighting
        const ambientLight = new THREE.AmbientLight(0x404040, 1);
        scene.add(ambientLight);
        const directionalLight = new THREE.DirectionalLight(0xffffff, 1);
        directionalLight.position.set(5, 5, 5);
        scene.add(directionalLight);

        // Camera
        camera.position.set(0, 0, 20);
        camera.lookAt(scene.position);

        // Controls
        const controls = {
            rotation: true,
            elevation: 1.0
        };

        const gui = new dat.GUI({ autoPlace: false });
        container.appendChild(gui.domElement);
        gui.add(controls, 'rotation').name('Auto Rotate');
        gui.add(controls, 'elevation', 0.5, 2).name('Terrain Height');

        function animate() {
            requestAnimationFrame(animate);
            if (controls.rotation) {
                pakistan.rotation.y += 0.005;
            }
            pakistan.scale.z = controls.elevation;
            renderer.render(scene, camera);
        }
        animate();

        // Handle window resize
        window.addEventListener('resize', () => {
            camera.aspect = container.clientWidth / container.clientHeight;
            camera.updateProjectionMatrix();
            renderer.setSize(container.clientWidth, container.clientHeight);
        });
    </script>
    """
    components.html(threejs_html, height=600)

# Update the show_disaster_monitor function
def show_disaster_monitor(data_collector):
    st.header("Advanced Disaster Monitoring System 🚨")
    
    earthquake_data = data_collector.fetch_usgs_earthquake_data()
    
    if earthquake_data:
        st.subheader("3D Terrain Visualization")
        create_3d_visualization(earthquake_data)  # Using the new visualization function
        
        # ... rest of your existing show_disaster_monitor function code ...

# The rest of your existing code remains the same, but make sure to remove any references 
# to create_cesium_component and replace them with create_3d_visualization
        for eq in earthquake_data['features']:
            coords = eq['geometry']['coordinates']
            mag = eq['properties']['mag']
            cesium_html += f"""
            earthquakeEntities.entities.add({{
                position: Cesium.Cartesian3.fromDegrees({coords[0]}, {coords[1]}, {coords[2]}),
                point: {{
                    pixelSize: {mag * 5},
                    color: Cesium.Color.RED.withAlpha(0.8),
                    outlineColor: Cesium.Color.WHITE,
                    outlineWidth: 2
                }},
                description: `Magnitude: {mag}<br>Depth: {coords[2]} km`
            }});
            """
    
    cesium_html += """
        viewer.camera.flyTo({
            destination: Cesium.Cartesian3.fromDegrees(69.3451, 30.3753, 1000000.0),
            orientation: {
                heading: Cesium.Math.toRadians(0.0),
                pitch: Cesium.Math.toRadians(-45.0),
                roll: 0.0
            }
        });
    </script>
    """
    components.html(cesium_html, height=600)

def train_weather_model(features_df, city):
    """Train ML model for weather predictions"""
    city_data = features_df[features_df['city'] == city].copy()
    
    # Prepare features
    feature_cols = ['hour', 'day', 'month', 'day_of_week', 
                   'temp_lag_1', 'temp_lag_24', 
                   'temp_rolling_mean_6h', 'temp_rolling_mean_24h']
    X = city_data[feature_cols]
    y = city_data['temperature']
    
    # Split data
    split_idx = int(len(X) * 0.8)
    X_train, X_test = X[:split_idx], X[split_idx:]
    y_train, y_test = y[:split_idx], y[split_idx:]
    
    # Train model
    model = XGBRegressor(n_estimators=100)
    model.fit(X_train, y_train)
    
    return model, X_test, y_test

def show_weather_analysis(data_collector):
    st.header("Advanced Weather Analysis 🌤️")
    
    weather_data = data_collector.fetch_weather_data()
    if weather_data:
        selected_city = st.selectbox(
            "Select City",
            options=[data['city'] for data in weather_data]
        )
        
        city_data = next(data for data in weather_data if data['city'] == selected_city)
        
        # Add download button for data
        st.markdown(download_csv(city_data['hourly'], f"{selected_city}_weather_data"), unsafe_allow_html=True)
        
        tabs = st.tabs(["Temperature Analysis", "Precipitation Insights", 
                       "Wind Patterns", "Humidity Trends", "ML Predictions"])
        
        with tabs[0]:
            col1, col2 = st.columns(2)
            with col1:
                # Enhanced temperature visualization
                fig = go.Figure()
                fig.add_trace(go.Scatter(
                    x=city_data['hourly']['datetime'],
                    y=city_data['hourly']['temperature'],
                    name='Temperature',
                    line=dict(color='red', width=2)
                ))
                fig.update_layout(
                    title='Temperature Trend with Range',
                    template='plotly_dark',
                    hovermode='x unified'
                )
                st.plotly_chart(fig, use_container_width=True)
            
            with col2:
                # Temperature distribution
                fig = px.histogram(
                    city_data['hourly'],
                    x='temperature',
                    nbins=30,
                    title='Temperature Distribution'
                )
                st.plotly_chart(fig, use_container_width=True)
        
        with tabs[1]:
            # Enhanced precipitation analysis
            col1, col2 = st.columns(2)
            with col1:
                fig = px.bar(
                    city_data['daily'],
                    x='date',
                    y='precipitation_sum',
                    title='Daily Precipitation',
                    color='precipitation_sum',
                    color_continuous_scale='Blues'
                )
                st.plotly_chart(fig, use_container_width=True)
            
            with col2:
                # Precipitation probability calculation
                precip_prob = (city_data['hourly']['precipitation'] > 0).mean() * 100
                st.metric(
                    "Precipitation Probability",
                    f"{precip_prob:.1f}%",
                    delta=f"{precip_prob - 50:.1f}%"
                )
        
        with tabs[2]:
            # Enhanced wind analysis
            fig = go.Figure()
            fig.add_trace(go.Scatter(
                x=city_data['hourly']['datetime'],
                y=city_data['hourly']['wind_speed'],
                name='Wind Speed',
                line=dict(color='blue', width=2)
            ))
            fig.add_trace(go.Scatter(
                x=city_data['hourly']['datetime'],
                y=city_data['hourly']['wind_speed'].rolling(24).mean(),
                name='24h Moving Average',
                line=dict(color='red', width=2, dash='dash')
            ))
            fig.update_layout(
                title='Wind Speed Analysis',
                template='plotly_dark',
                hovermode='x unified'
            )
            st.plotly_chart(fig, use_container_width=True)
        
        with tabs[3]:
            # Enhanced humidity analysis
            col1, col2 = st.columns(2)
            with col1:
                fig = px.line(
                    city_data['hourly'],
                    x='datetime',
                    y='humidity',
                    title='Humidity Trends',
                    color_discrete_sequence=['green']
                )
                st.plotly_chart(fig, use_container_width=True)
            
            with col2:
                # Humidity comfort zones
                comfort_zones = pd.cut(
                    city_data['hourly']['humidity'],
                    bins=[0, 30, 45, 60, 100],
                    labels=['Dry', 'Comfortable', 'Moderate', 'Humid']
                ).value_counts()
                fig = px.pie(
                    values=comfort_zones.values,
                    names=comfort_zones.index,
                    title='Humidity Comfort Zones'
                )
                st.plotly_chart(fig, use_container_width=True)
        
        with tabs[4]:
            st.subheader("Machine Learning Weather Predictions")
            
            # Prepare data for ML
            features_df = data_collector.create_ml_features(weather_data)
            model, X_test, y_test = train_weather_model(features_df, selected_city)
            
            # Make predictions
            predictions = model.predict(X_test)
            
            # Show predictions vs actual
            fig = go.Figure()
            fig.add_trace(go.Scatter(
                x=X_test.index,
                y=y_test,
                name='Actual Temperature',
                line=dict(color='blue')
            ))
            fig.add_trace(go.Scatter(
                x=X_test.index,
                y=predictions,
                name='Predicted Temperature',
                line=dict(color='red', dash='dash')
            ))
            fig.update_layout(
                title='Temperature Predictions vs Actual',
                template='plotly_dark',
                hovermode='x unified'
            )
            st.plotly_chart(fig, use_container_width=True)
            
            # Model metrics
            mae = np.mean(np.abs(predictions - y_test))
            rmse = np.sqrt(np.mean((predictions - y_test)**2))
            
            col1, col2 = st.columns(2)
            col1.metric("Mean Absolute Error", f"{mae:.2f}°C")
            col2.metric("Root Mean Square Error", f"{rmse:.2f}°C")

def show_disaster_monitor(data_collector):
    st.header("Advanced Disaster Monitoring System 🚨")
    
    earthquake_data = data_collector.fetch_usgs_earthquake_data()
    
    if earthquake_data:
        # Enhanced 3D visualization
        st.subheader("3D Terrain Analysis")
        create_3d_visualization(earthquake_data)
        
        # Advanced earthquake analysis
        st.subheader("Earthquake Analysis Dashboard")
        
        # Create DataFrame for analysis
        eq_df = pd.DataFrame([
            {
                'time': datetime.fromtimestamp(eq['properties']['time']/1000),
                'magnitude': eq['properties']['mag'],
                'location': eq['properties']['place'],
                'depth': eq['geometry']['coordinates'][2],
                'lat': eq['geometry']['coordinates'][1],
                'lon': eq['geometry']['coordinates'][0]
            }
            for eq in earthquake_data['features']
        ])
        
        col1, col2 = st.columns(2)
        
        with col1:
            # Magnitude distribution
            fig = px.histogram(
                eq_df,
                x='magnitude',
                nbins=20,
                title='Earthquake Magnitude Distribution',
                color_discrete_sequence=['red']
            )
            st.plotly_chart(fig, use_container_width=True)
        
        with col2:
            # Depth vs Magnitude scatter
            fig = px.scatter(
                eq_df,
                x='depth',
                y='magnitude',
                title='Depth vs Magnitude',
                color='magnitude',
                size='magnitude',
                color_continuous_scale='Viridis'
            )
            st.plotly_chart(fig, use_container_width=True)
        
        # Time series analysis
        st.subheader("Temporal Analysis")
        eq_df# Time series analysis
        eq_df['date'] = eq_df['time'].dt.date
        daily_counts = eq_df.groupby('date').size().reset_index(name='count')
        
        fig = px.line(
            daily_counts,
            x='date',
            y='count',
            title='Daily Earthquake Frequency',
            line_shape='spline'
        )
        st.plotly_chart(fig, use_container_width=True)
        
        # Risk assessment
        st.subheader("Seismic Risk Assessment")
        risk_zones = folium.Map(location=[30.3753, 69.3451], zoom_start=5)
        
        # Create heatmap layer
        heat_data = [[row['lat'], row['lon'], row['magnitude']] for _, row in eq_df.iterrows()]
        folium.plugins.HeatMap(heat_data).add_to(risk_zones)
        
        # Add fault lines (simplified example)
        fault_lines = {
            'Main Boundary Thrust': [[34.0151, 71.5249], [33.7294, 73.0931]],
            'Chaman Fault': [[30.1798, 66.9750], [25.1216, 62.3254]],
        }
        
        for name, coords in fault_lines.items():
            folium.PolyLine(
                coords,
                color='red',
                weight=2,
                popup=name
            ).add_to(risk_zones)
        
        folium_static(risk_zones)
        
        # Earthquake prediction model
        st.subheader("Seismic Activity Prediction")
        
        # Prepare time series data for prediction
        daily_counts['ds'] = pd.to_datetime(daily_counts['date'])
        daily_counts['y'] = daily_counts['count']
        
        # Train Prophet model
        model = Prophet(yearly_seasonality=True, weekly_seasonality=True)
        model.fit(daily_counts[['ds', 'y']])
        
        # Make future predictions
        future_dates = model.make_future_dataframe(periods=30)
        forecast = model.predict(future_dates)
        
        # Plot predictions
        fig = go.Figure()
        fig.add_trace(go.Scatter(
            x=daily_counts['ds'],
            y=daily_counts['y'],
            name='Actual',
            line=dict(color='blue')
        ))
        fig.add_trace(go.Scatter(
            x=forecast['ds'],
            y=forecast['yhat'],
            name='Predicted',
            line=dict(color='red', dash='dash')
        ))
        fig.add_trace(go.Scatter(
            x=forecast['ds'],
            y=forecast['yhat_upper'],
            fill=None,
            mode='lines',
            line=dict(color='rgba(255,0,0,0)'),
            showlegend=False
        ))
        fig.add_trace(go.Scatter(
            x=forecast['ds'],
            y=forecast['yhat_lower'],
            fill='tonexty',
            mode='lines',
            line=dict(color='rgba(255,0,0,0)'),
            name='Prediction Interval'
        ))
        fig.update_layout(
            title='Seismic Activity Forecast (30 Days)',
            xaxis_title='Date',
            yaxis_title='Number of Earthquakes',
            template='plotly_dark'
        )
        st.plotly_chart(fig, use_container_width=True)

def show_environmental_data(data_collector):
    st.header("Advanced Environmental Analysis 🌿")
    
    aqi_data = data_collector.fetch_air_quality_data()
    
    if aqi_data is not None:
        selected_city = st.selectbox("Select City", aqi_data['city'].unique())
        city_data = aqi_data[aqi_data['city'] == selected_city].copy()
        
        # Add download button
        st.markdown(download_csv(city_data, f"{selected_city}_air_quality_data"), unsafe_allow_html=True)
        
        # Enhanced AQI calculation with weights
        weights = {
            'PM2.5': 0.3,
            'PM10': 0.2,
            'NO2': 0.2,
            'O3': 0.2,
            'CO': 0.1
        }
        
        # Normalize and calculate weighted AQI
        for pollutant in weights.keys():
            max_val = city_data[pollutant].max()
            city_data[f'{pollutant}_normalized'] = city_data[pollutant] / max_val * 100
            city_data[f'{pollutant}_weighted'] = city_data[f'{pollutant}_normalized'] * weights[pollutant]
        
        city_data['AQI'] = sum(city_data[f'{p}_weighted'] for p in weights.keys())
        
        tabs = st.tabs(["AQI Dashboard", "Pollutant Analysis", "Trends & Forecasting", "Health Impact"])
        
        with tabs[0]:
            col1, col2, col3 = st.columns(3)
            
            current_aqi = city_data['AQI'].iloc[-1]
            with col1:
                st.metric(
                    "Current AQI",
                    f"{current_aqi:.1f}",
                    delta=f"{current_aqi - city_data['AQI'].iloc[-2]:.1f}"
                )
            
            # AQI categories
            aqi_categories = pd.cut(
                city_data['AQI'],
                bins=[0, 50, 100, 150, 200, 300, float('inf')],
                labels=['Good', 'Moderate', 'Unhealthy for Sensitive Groups', 'Unhealthy', 'Very Unhealthy', 'Hazardous']
            ).value_counts()
            
            with col2:
                fig = px.pie(
                    values=aqi_categories.values,
                    names=aqi_categories.index,
                    title='AQI Distribution'
                )
                st.plotly_chart(fig, use_container_width=True)
            
            with col3:
                # Daily pattern
                hourly_avg = city_data.groupby(city_data['datetime'].dt.hour)['AQI'].mean()
                fig = px.line(
                    x=hourly_avg.index,
                    y=hourly_avg.values,
                    title='Daily AQI Pattern',
                    labels={'x': 'Hour of Day', 'y': 'Average AQI'}
                )
                st.plotly_chart(fig, use_container_width=True)
        
        with tabs[1]:
            # Pollutant correlation analysis
            pollutants = ['PM2.5', 'PM10', 'CO', 'NO2', 'O3']
            corr_matrix = city_data[pollutants].corr()
            
            fig = px.imshow(
                corr_matrix,
                title='Pollutant Correlation Matrix',
                color_continuous_scale='RdBu'
            )
            st.plotly_chart(fig, use_container_width=True)
            
            # Individual pollutant analysis
            selected_pollutant = st.selectbox("Select Pollutant", pollutants)
            
            col1, col2 = st.columns(2)
            with col1:
                fig = px.line(
                    city_data,
                    x='datetime',
                    y=selected_pollutant,
                    title=f'{selected_pollutant} Trend'
                )
                st.plotly_chart(fig, use_container_width=True)
            
            with col2:
                fig = px.box(
                    city_data,
                    y=selected_pollutant,
                    title=f'{selected_pollutant} Distribution'
                )
                st.plotly_chart(fig, use_container_width=True)
        
        with tabs[2]:
            # Time series decomposition
            from statsmodels.tsa.seasonal import seasonal_decompose
            
            # Resample to hourly data for decomposition
            hourly_data = city_data.set_index('datetime')['AQI'].resample('H').mean()
            decomposition = seasonal_decompose(hourly_data, period=24)
            
            fig = make_subplots(rows=4, cols=1, subplot_titles=('Observed', 'Trend', 'Seasonal', 'Residual'))
            fig.add_trace(go.Scatter(x=hourly_data.index, y=hourly_data.values, name='Observed'), row=1, col=1)
            fig.add_trace(go.Scatter(x=hourly_data.index, y=decomposition.trend, name='Trend'), row=2, col=1)
            fig.add_trace(go.Scatter(x=hourly_data.index, y=decomposition.seasonal, name='Seasonal'), row=3, col=1)
            fig.add_trace(go.Scatter(x=hourly_data.index, y=decomposition.resid, name='Residual'), row=4, col=1)
            fig.update_layout(height=800, title_text="AQI Time Series Decomposition")
            st.plotly_chart(fig, use_container_width=True)
        
        with tabs[3]:
            st.subheader("Health Impact Assessment")
            
            # Define health impact thresholds
            impact_thresholds = {
                'PM2.5': [12, 35.4, 55.4, 150.4],
                'PM10': [54, 154, 254, 354],
                'NO2': [53, 100, 360, 649],
                'O3': [54, 70, 85, 105],
                'CO': [4.4, 9.4, 12.4, 15.4]
            }
            
            # Calculate current health risks
            current_risks = {}
            for pollutant, thresholds in impact_thresholds.items():
                current_val = city_data[pollutant].iloc[-1]
                if current_val <= thresholds[0]:
                    risk = 'Low'
                elif current_val <= thresholds[1]:
                    risk = 'Moderate'
                elif current_val <= thresholds[2]:
                    risk = 'High'
                else:
                    risk = 'Very High'
                current_risks[pollutant] = {'value': current_val, 'risk': risk}
            
            # Display health risks
            col1, col2 = st.columns(2)
            with col1:
                for pollutant, data in current_risks.items():
                    st.metric(
                        f"{pollutant} Health Risk",
                        data['risk'],
                        f"{data['value']:.1f}"
                    )
            
            with col2:
                # Health recommendations based on current AQI
                if current_aqi <= 50:
                    st.success("Air quality is good. Outdoor activities are safe.")
                elif current_aqi <= 100:
                    st.info("Sensitive individuals should consider reducing prolonged outdoor exertion.")
                elif current_aqi <= 150:
                    st.warning("Everyone should reduce prolonged outdoor exertion.")
                else:
                    st.error("Avoid outdoor activities. Use air purifiers indoors.")

def main():
    st.title("🌍 Pakistan Climate & Disaster Monitoring System")
    
    # Add dashboard info
    st.sidebar.image("https://upload.wikimedia.org/wikipedia/commons/3/32/Flag_of_Pakistan.svg", width=100)
    st.sidebar.title("Dashboard Controls")
    
    data_collector = DataCollector()
    
    # Enhanced navigation
    page = st.sidebar.radio(
        "Select Module",
        ["Weather Analysis", "Disaster Monitor", "Environmental Data"],
        format_func=lambda x: f"📊 {x}" if x == "Weather Analysis" else 
                            f"🚨 {x}" if x == "Disaster Monitor" else 
                            f"🌿 {x}"
    )
    
    # Add data timestamp
    st.sidebar.markdown("---")
    st.sidebar.markdown(f"Last updated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
    
    if page == "Weather Analysis":
        show_weather_analysis(data_collector)
    elif page == "Disaster Monitor":
        show_disaster_monitor(data_collector)
    elif page == "Environmental Data":
        show_environmental_data(data_collector)
    st.markdown("---")
    st.markdown("""
        <div style='text-align: center'>
            
            <p>Created by Muhammad Shaheer</p>
            
        </div>
    """, unsafe_allow_html=True)

if __name__ == "__main__":
    main()