app.py

import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
import keras
from keras.models import Sequential
from keras.layers import Dense
from sklearn.metrics import confusion_matrix, classification_report
import plotly.express as px
import plotly.graph_objects as go

def load_and_prep_data():
    # Define column names
    HDNames = ['age', 'sex', 'cp', 'trestbps', 'chol', 'fbs', 'restecg', 'thalach', 
               'exang', 'oldpeak', 'slope', 'ca', 'hal', 'HeartDisease']
    
    # Load data - Update path to where your Excel file is stored
    data = pd.read_excel('Ch3.ClevelandData.xlsx', names=HDNames)
    
    # Handle missing data
    data_new = data.replace("?", np.nan)
    data_new = data_new.dropna()
    
    return data_new

def scale_and_split_data(data_new):
    feature_names = ['age', 'sex', 'cp', 'trestbps', 'chol', 'fbs', 'restecg', 
                    'thalach', 'exang', 'oldpeak', 'slope', 'ca', 'hal']
    
    features = pd.DataFrame(data_new.iloc[:, 0:13], columns=feature_names)
    target = pd.DataFrame(data_new.iloc[:,13], columns=["HeartDisease"])
    
    scaler = StandardScaler()
    FeatureScaled = scaler.fit_transform(features)
    FeatureScaled = pd.DataFrame(FeatureScaled, columns=feature_names)
    
    X_train, X_test, y_train, y_test = train_test_split(
        FeatureScaled, target, test_size=0.30, random_state=5
    )
    
    return X_train, X_test, y_train, y_test, FeatureScaled, scaler, features

def create_model():
    model = Sequential()
    model.add(Dense(30, input_dim=13, activation="tanh"))
    model.add(Dense(20, activation="tanh"))
    model.add(Dense(1, activation="sigmoid"))
    
    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
    return model

def plot_correlation_heatmap(FeatureScaled):
    corrMatrix = FeatureScaled.corr()
    fig = px.imshow(corrMatrix, 
                    labels=dict(color="Correlation"),
                    title="Feature Correlation Heatmap")
    return fig

def plot_feature_distributions(FeatureScaled):
    fig = go.Figure()
    for column in FeatureScaled.columns:
        fig.add_trace(go.Box(y=FeatureScaled[column], name=column))
    fig.update_layout(title="Feature Distributions",
                     showlegend=False,
                     height=600)
    return fig

def make_prediction(model, scaler, features, input_data):
    # Transform input data using the same scaler
    scaled_input = scaler.transform([input_data])
    # Make prediction
    prediction = model.predict(scaled_input)
    return prediction[0][0]

def main():
    st.set_page_config(page_title="Heart Disease Classification", layout="wide")
    
    st.title("Heart Disease Classification")
    st.markdown("""
    This application predicts the likelihood of heart disease based on various medical parameters.
    The model uses neural networks to make predictions with an accuracy of approximately 82%.
    """)
    
    # Load and prepare data
    with st.spinner("Loading and preparing data..."):
        data_new = load_and_prep_data()
        X_train, X_test, y_train, y_test, FeatureScaled, scaler, features = scale_and_split_data(data_new)
    
    # Training section
    with st.expander("Model Training"):
        if st.button("Train Model"):
            with st.spinner("Training model..."):
                model = create_model()
                history = model.fit(X_train, y_train, epochs=1000, verbose=0)
                score = model.evaluate(X_test, y_test, verbose=0)
                st.success(f"Model trained! Accuracy: {score[1]:.2f}")
                st.session_state['model'] = model
                st.session_state['scaler'] = scaler
    
    # Data visualization section
    st.header("Data Visualization")
    viz_option = st.selectbox(
        "Choose visualization",
        ["Correlation Heatmap", "Feature Distributions"]
    )
    
    if viz_option == "Correlation Heatmap":
        st.plotly_chart(plot_correlation_heatmap(FeatureScaled), use_container_width=True)
    else:
        st.plotly_chart(plot_feature_distributions(FeatureScaled), use_container_width=True)
    
    # Prediction section
    st.header("Make Prediction")
    
    col1, col2, col3 = st.columns(3)
    
    with col1:
        age = st.number_input("Age", min_value=20, max_value=100, value=50)
        sex = st.selectbox("Sex", [0, 1], format_func=lambda x: "Female" if x == 0 else "Male")
        cp = st.selectbox("Chest Pain Type", [1, 2, 3, 4])
        trestbps = st.number_input("Resting Blood Pressure (mmHg)", 90, 200, 120)
        chol = st.number_input("Cholesterol (mg/dl)", 100, 600, 250)
    
    with col2:
        fbs = st.selectbox("Fasting Blood Sugar > 120 mg/dl", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
        restecg = st.selectbox("Resting ECG Results", [0, 1, 2])
        thalach = st.number_input("Maximum Heart Rate", 70, 220, 150)
        exang = st.selectbox("Exercise Induced Angina", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
    
    with col3:
        oldpeak = st.number_input("ST Depression", 0.0, 10.0, 1.0)
        slope = st.selectbox("ST Slope", [1, 2, 3])
        ca = st.selectbox("Number of Major Vessels", [0, 1, 2, 3])
        hal = st.selectbox("Thalassemia", [3, 6, 7])
    
    if st.button("Predict"):
        if 'model' not in st.session_state:
            st.error("Please train the model first!")
        else:
            input_data = [age, sex, cp, trestbps, chol, fbs, restecg, thalach, 
                         exang, oldpeak, slope, ca, hal]
            prediction = make_prediction(st.session_state['model'], 
                                      st.session_state['scaler'], 
                                      features, 
                                      input_data)
            
            risk_level = "High Risk" if prediction > 0.5 else "Low Risk"
            risk_color = "red" if prediction > 0.5 else "green"
            
            st.markdown(f"""
            <h2 style='text-align: center; color: {risk_color};'>
                {risk_level}
            </h2>
            <p style='text-align: center;'>
                Probability of heart disease: {prediction:.2%}
            </p>
            """, unsafe_allow_html=True)

if __name__ == "__main__":
    main()