Update app.py

app.py

@@ -1,13 +1,13 @@
 import streamlit as st
 import pandas as pd
 import numpy as np
-from sklearn.ensemble import RandomForestClassifier
 from sklearn.model_selection import train_test_split
+from sklearn.ensemble import RandomForestClassifier
 from sklearn.metrics import accuracy_score
-import pickle
 import matplotlib.pyplot as plt
 import seaborn as sns
 from PIL import Image
+import pickle
 
 # Set page configuration
 st.set_page_config(
@@ -17,281 +17,196 @@ st.set_page_config(
     initial_sidebar_state="expanded"
 )
 
-# Custom CSS for styling with new color scheme
+# Custom CSS for styling
 st.markdown("""
     <style>
     .main {
-        background-color: #f5f9ff;
+        background-color: #f0f2f6;
     }
     .stButton>button {
-        background: linear-gradient(135deg, #6e8efb, #a777e3);
+        background-color: #4CAF50;
         color: white;
-        border-radius: 8px;
+        border-radius: 5px;
         border: none;
-        padding: 12px 28px;
+        padding: 10px 24px;
         text-align: center;
         text-decoration: none;
         display: inline-block;
         font-size: 16px;
-        margin: 8px 2px;
+        margin: 4px 2px;
         cursor: pointer;
-        transition: all 0.3s;
-        box-shadow: 0 4px 6px rgba(0,0,0,0.1);
     }
     .stButton>button:hover {
-        background: linear-gradient(135deg, #5d7de8, #9665d6);
-        box-shadow: 0 6px 10px rgba(0,0,0,0.15);
-        transform: translateY(-2px);
+        background-color: #45a049;
     }
     .stSelectbox, .stNumberInput, .stSlider {
         background-color: white;
-        border-radius: 8px;
-        padding: 12px;
-        box-shadow: 0 2px 4px rgba(0,0,0,0.05);
+        border-radius: 5px;
     }
     .css-1aumxhk {
-        background-color: #ffffff;
-        background-image: none;
-        color: #2c3e50;
-    }
-    .reportview-container .main .block-container {
-        padding-top: 2rem;
-        padding-bottom: 2rem;
-    }
-    .header {
-        font-size: 2.5em;
-        color: #2c3e50;
-        text-align: center;
-        margin-bottom: 0.5em;
-        background: linear-gradient(135deg, #4b6cb7, #182848);
-        -webkit-background-clip: text;
-        -webkit-text-fill-color: transparent;
-        font-weight: 700;
+        background-color: #e8f4fc;
+        border-radius: 10px;
+        padding: 20px;
     }
-    .subheader {
-        font-size: 1.2em;
-        color: #5d6d7e;
+    .title {
+        color: #1e3d6b;
         text-align: center;
-        margin-bottom: 2em;
-    }
-    .sidebar .sidebar-content {
-        background: linear-gradient(180deg, #e0f7fa, #b2ebf2);
-        border-right: 1px solid #b2ebf2;
-    }
-    .st-bb {
-        background-color: white;
-    }
-    .st-at {
-        background-color: #4b6cb7;
-    }
-    .st-ae {
-        background-color: #f5f9ff;
-    }
-    .st-cg {
-        color: #4b6cb7;
-    }
-    .st-cn {
-        background-color: #4b6cb7;
-    }
-    .stTab {
-        background-color: #f5f9ff;
-        border-radius: 8px;
-        padding: 10px;
-    }
-    .stTab > div > div {
-        background: linear-gradient(135deg, #e0f7fa, #b2ebf2);
-        border-radius: 8px;
     }
     </style>
-""", unsafe_allow_html=True)
-
-# App header
-st.markdown('<p class="header">🌧️ Rainfall Prediction Model</p>', unsafe_allow_html=True)
-st.markdown('<p class="subheader">Predict whether it will rain tomorrow based on weather data</p>', unsafe_allow_html=True)
-
-# Load or train model
-@st.cache_resource
-def load_model():
-    try:
-        # Try to load pre-trained model
-        model = pickle.load(open('rain_prediction_model.pkl', 'rb'))
-        return model
-    except:
-        # If no model exists, train a new one (for demo purposes)
-        # This would normally be done separately
-        from sklearn.datasets import make_classification
-        X, y = make_classification(n_samples=1000, n_features=10, random_state=42)
-        model = RandomForestClassifier(n_estimators=100, random_state=42)
-        model.fit(X, y)
-        return model
+    """, unsafe_allow_html=True)
 
-model = load_model()
+# Title and description
+st.title("🌧️ Rainfall Prediction App")
+st.markdown("""
+    Predict the probability of rainfall tomorrow based on weather data.
+    This machine learning model uses Random Forest Classifier for accurate predictions.
+    """)
 
-# Sample data (in a real app, you would load your actual dataset)
+# Load sample data or create synthetic data
 @st.cache_data
 def load_data():
+    # In a real app, you would load your actual dataset here
+    # This is synthetic data for demonstration
     data = {
-        'MinTemp': np.random.normal(12, 5, 100),
-        'MaxTemp': np.random.normal(24, 5, 100),
-        'Rainfall': np.random.exponential(2, 100),
-        'Evaporation': np.random.normal(5, 2, 100),
-        'Sunshine': np.random.normal(8, 3, 100),
-        'WindGustSpeed': np.random.randint(30, 100, 100),
-        'Humidity9am': np.random.randint(40, 100, 100),
-        'Humidity3pm': np.random.randint(30, 90, 100),
-        'Pressure9am': np.random.normal(1015, 5, 100),
-        'Pressure3pm': np.random.normal(1013, 5, 100),
-        'RainToday': np.random.choice([0, 1], 100, p=[0.7, 0.3]),
-        'RainTomorrow': np.random.choice([0, 1], 100, p=[0.7, 0.3])
+        'MinTemp': np.random.uniform(5, 25, 1000),
+        'MaxTemp': np.random.uniform(15, 40, 1000),
+        'Rainfall': np.random.uniform(0, 50, 1000),
+        'Evaporation': np.random.uniform(0, 20, 1000),
+        'Sunshine': np.random.uniform(0, 14, 1000),
+        'Humidity9am': np.random.randint(20, 100, 1000),
+        'Humidity3pm': np.random.randint(20, 100, 1000),
+        'Pressure9am': np.random.uniform(990, 1040, 1000),
+        'Pressure3pm': np.random.uniform(990, 1040, 1000),
+        'Cloud9am': np.random.randint(0, 9, 1000),
+        'Cloud3pm': np.random.randint(0, 9, 1000),
+        'Temp9am': np.random.uniform(10, 35, 1000),
+        'Temp3pm': np.random.uniform(15, 40, 1000),
+        'RainToday': np.random.choice([0, 1], 1000, p=[0.7, 0.3]),
+        'RainTomorrow': np.random.choice([0, 1], 1000, p=[0.7, 0.3])
     }
     return pd.DataFrame(data)
 
 df = load_data()
 
-# Sidebar for navigation
-with st.sidebar:
-    st.title("Navigation")
-    app_mode = st.selectbox("Choose a page", ["Prediction", "Data Exploration", "About"])
-    
-    st.markdown("---")
-    st.markdown("### Weather Parameters")
-    st.markdown("Adjust the sliders to set weather conditions for prediction.")
+# Sidebar for user input
+st.sidebar.header("User Input Features")
 
-# Main app content
-if app_mode == "Prediction":
-    col1, col2 = st.columns([1, 1])
-    
-    with col1:
-        st.subheader("Enter Weather Data")
-        
-        min_temp = st.slider("Minimum Temperature (°C)", -5.0, 30.0, 15.0)
-        max_temp = st.slider("Maximum Temperature (°C)", 5.0, 45.0, 25.0)
-        rainfall = st.slider("Rainfall (mm)", 0.0, 50.0, 0.0)
-        evaporation = st.slider("Evaporation (mm)", 0.0, 20.0, 5.0)
-        
-    with col2:
-        st.subheader("Additional Parameters")
-        
-        sunshine = st.slider("Sunshine (hours)", 0.0, 14.0, 8.0)
-        wind_gust = st.slider("Wind Gust Speed (km/h)", 30, 100, 50)
-        humidity_9am = st.slider("Humidity at 9am (%)", 0, 100, 70)
-        humidity_3pm = st.slider("Humidity at 3pm (%)", 0, 100, 50)
-        pressure_9am = st.slider("Pressure at 9am (hPa)", 990.0, 1040.0, 1015.0)
-        pressure_3pm = st.slider("Pressure at 3pm (hPa)", 990.0, 1040.0, 1013.0)
-    
-    # Prediction button
-    if st.button("Predict Rainfall Tomorrow"):
-        # Prepare input data
-        input_data = np.array([[min_temp, max_temp, rainfall, evaporation, sunshine, 
-                               wind_gust, humidity_9am, humidity_3pm, 
-                               pressure_9am, pressure_3pm]])
-        
-        # Make prediction
-        prediction = model.predict(input_data)
-        prediction_proba = model.predict_proba(input_data)
-        
-        # Display results
-        st.markdown("---")
-        st.subheader("Prediction Results")
-        
-        col_result1, col_result2 = st.columns([1, 1])
-        
-        with col_result1:
-            if prediction[0] == 1:
-                st.markdown("### 🌧️ Result: It will rain tomorrow!")
-                st.image("https://cdn-icons-png.flaticon.com/512/4150/4150904.png", width=150)
-            else:
-                st.markdown("### ☀️ Result: No rain expected tomorrow!")
-                st.image("https://cdn-icons-png.flaticon.com/512/3222/3222807.png", width=150)
-        
-        with col_result2:
-            st.markdown("### Probability")
-            st.write(f"Probability of rain: {prediction_proba[0][1]*100:.2f}%")
-            st.write(f"Probability of no rain: {prediction_proba[0][0]*100:.2f}%")
-            
-            # Create a gauge chart with new colors
-            fig, ax = plt.subplots(figsize=(6, 1))
-            ax.barh(['Rain Probability'], [prediction_proba[0][1]], color='#6e8efb')
-            ax.set_xlim(0, 1)
-            ax.set_xticks([])
-            ax.text(0.5, 0, f"{prediction_proba[0][1]*100:.1f}%", 
-                   ha='center', va='center', color='white', fontsize=12,
-                   bbox=dict(facecolor='rgba(0,0,0,0.2)', edgecolor='none', boxstyle='round,pad=0.2'))
-            st.pyplot(fig)
+# Function to get user input
+def user_input_features():
+    min_temp = st.sidebar.slider('Minimum Temperature (°C)', 0.0, 30.0, 15.0)
+    max_temp = st.sidebar.slider('Maximum Temperature (°C)', 10.0, 45.0, 25.0)
+    rainfall = st.sidebar.slider('Rainfall (mm)', 0.0, 100.0, 0.0)
+    humidity_9am = st.sidebar.slider('Humidity at 9am (%)', 0, 100, 60)
+    humidity_3pm = st.sidebar.slider('Humidity at 3pm (%)', 0, 100, 50)
+    pressure_9am = st.sidebar.slider('Pressure at 9am (hPa)', 990.0, 1040.0, 1015.0)
+    pressure_3pm = st.sidebar.slider('Pressure at 3pm (hPa)', 990.0, 1040.0, 1015.0)
+    cloud_9am = st.sidebar.slider('Cloud cover at 9am (oktas)', 0, 8, 3)
+    cloud_3pm = st.sidebar.slider('Cloud cover at 3pm (oktas)', 0, 8, 4)
+    rain_today = st.sidebar.selectbox('Rain Today?', ('No', 'Yes'))
     
-elif app_mode == "Data Exploration":
-    st.subheader("Weather Data Exploration")
-    
-    tab1, tab2, tab3 = st.tabs(["Data Overview", "Visualizations", "Statistics"])
-    
-    with tab1:
-        st.write("### Sample Data")
-        st.dataframe(df.head(10))
-        
-        st.write("### Data Description")
-        st.write(df.describe())
-    
-    with tab2:
-        st.write("### Temperature Distribution")
-        fig, ax = plt.subplots()
-        sns.histplot(df['MaxTemp'], kde=True, ax=ax, color='#4b6cb7')
-        ax.set_facecolor('#f5f9ff')
-        fig.patch.set_facecolor('#f5f9ff')
-        st.pyplot(fig)
-        
-        st.write("### Rainfall vs. Humidity")
-        fig, ax = plt.subplots()
-        sns.scatterplot(data=df, x='Humidity9am', y='Rainfall', hue='RainTomorrow', 
-                        palette=['#a777e3', '#6e8efb'], ax=ax)
-        ax.set_facecolor('#f5f9ff')
-        fig.patch.set_facecolor('#f5f9ff')
-        st.pyplot(fig)
+    data = {
+        'MinTemp': min_temp,
+        'MaxTemp': max_temp,
+        'Rainfall': rainfall,
+        'Humidity9am': humidity_9am,
+        'Humidity3pm': humidity_3pm,
+        'Pressure9am': pressure_9am,
+        'Pressure3pm': pressure_3pm,
+        'Cloud9am': cloud_9am,
+        'Cloud3pm': cloud_3pm,
+        'RainToday': 1 if rain_today == 'Yes' else 0
+    }
     
-    with tab3:
-        st.write("### Correlation Matrix")
-        fig, ax = plt.subplots(figsize=(10, 8))
-        sns.heatmap(df.corr(), annot=True, cmap='coolwarm', ax=ax, center=0,
-                   cbar_kws={'label': 'Correlation Coefficient'})
-        ax.set_facecolor('#f5f9ff')
-        fig.patch.set_facecolor('#f5f9ff')
-        st.pyplot(fig)
-        
-        st.write("### Rain Tomorrow Distribution")
-        fig, ax = plt.subplots()
-        df['RainTomorrow'].value_counts().plot(kind='bar', color=['#6e8efb', '#a777e3'], ax=ax)
-        ax.set_facecolor('#f5f9ff')
-        fig.patch.set_facecolor('#f5f9ff')
-        st.pyplot(fig)
+    features = pd.DataFrame(data, index=[0])
+    return features
+
+input_df = user_input_features()
 
-elif app_mode == "About":
-    st.subheader("About This App")
+# Main panel
+st.subheader("User Input Parameters")
+st.write(input_df)
+
+# Train model
+@st.cache_resource
+def train_model():
+    # Prepare data
+    X = df[['MinTemp', 'MaxTemp', 'Rainfall', 'Humidity9am', 'Humidity3pm', 
+            'Pressure9am', 'Pressure3pm', 'Cloud9am', 'Cloud3pm', 'RainToday']]
+    y = df['RainTomorrow']
     
-    st.markdown("""
-    This Rainfall Prediction App uses machine learning to predict whether it will rain tomorrow based on current weather conditions.
+    # Train-test split
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
     
-    ### Features:
-    - Interactive input sliders for weather parameters
-    - Real-time prediction with probability scores
-    - Data exploration and visualization tools
-    - Beautiful and responsive UI design
+    # Train model
+    model = RandomForestClassifier(n_estimators=100, random_state=42)
+    model.fit(X_train, y_train)
     
-    ### Model Information:
-    - Algorithm: Random Forest Classifier
-    - Input Features: Temperature, Humidity, Rainfall, Wind Speed, etc.
-    - Output: Probability of rain tomorrow
+    # Evaluate model
+    y_pred = model.predict(X_test)
+    accuracy = accuracy_score(y_test, y_pred)
     
-    ### How to Use:
-    1. Navigate to the Prediction page
-    2. Adjust the sliders to match current weather conditions
-    3. Click the "Predict" button to see the results
+    return model, accuracy
+
+model, accuracy = train_model()
+
+# Make predictions
+prediction = model.predict(input_df)
+prediction_proba = model.predict_proba(input_df)
+
+# Display results
+st.subheader('Prediction')
+rain_tomorrow = 'Yes' if prediction[0] == 1 else 'No'
+st.markdown(f"<h3 style='text-align: center; color: {'red' if rain_tomorrow == 'Yes' else 'green'};'>"
+            f"Will it rain tomorrow? {rain_tomorrow}</h3>", 
+            unsafe_allow_html=True)
+
+st.subheader('Prediction Probability')
+st.write(f"Probability of rain tomorrow: {prediction_proba[0][1]*100:.2f}%")
+st.write(f"Probability of no rain tomorrow: {prediction_proba[0][0]*100:.2f}%")
+
+# Show model accuracy
+st.sidebar.subheader('Model Accuracy')
+st.sidebar.write(f"Accuracy: {accuracy*100:.2f}%")
+
+# Visualization section
+st.subheader("Data Visualization")
+
+col1, col2 = st.columns(2)
+
+with col1:
+    st.write("**Temperature Distribution**")
+    fig, ax = plt.subplots()
+    sns.histplot(df['MaxTemp'], kde=True, ax=ax)
+    plt.xlabel('Maximum Temperature (°C)')
+    st.pyplot(fig)
+
+with col2:
+    st.write("**Humidity at 9am vs 3pm**")
+    fig, ax = plt.subplots()
+    sns.scatterplot(x='Humidity9am', y='Humidity3pm', hue='RainTomorrow', data=df, ax=ax)
+    plt.xlabel('Humidity at 9am (%)')
+    plt.ylabel('Humidity at 3pm (%)')
+    st.pyplot(fig)
+
+# Add some information
+st.markdown("""
+    ### About This App
+    This app predicts the probability of rainfall tomorrow using weather data.
+    The model was trained using Random Forest Classifier on historical weather data.
     
-    This app is designed for demonstration purposes and can be deployed on Hugging Face Spaces.
+    **Note:** This is a demonstration app using synthetic data. For real-world applications,
+    you should train the model with actual weather data from your region.
     """)
-    
-    st.markdown("---")
-    st.markdown("""
-    <div style="text-align: center; padding: 20px; background: linear-gradient(135deg, #e0f7fa, #b2ebf2); border-radius: 10px;">
-        <h3 style="color: #2c3e50;">Created with ❤️ using Streamlit</h3>
-        <p style="color: #5d6d7e;">Blue-Teal-Purple Gradient Theme</p>
-    </div>
-    """, unsafe_allow_html=True)
+
+# Add download button for sample data
+@st.cache_data
+def convert_df(df):
+    return df.to_csv().encode('utf-8')
+
+csv = convert_df(df.head(100))
+st.download_button(
+    label="Download Sample Data (CSV)",
+    data=csv,
+    file_name='sample_weather_data.csv',
+    mime='text/csv'
+)