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.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, confusion_matrix, roc_curve, auc

st.title("🩺 Diabetes Prediction App")

# Load dataset
@st.cache_data
def load_data():
    file_path = "diabetes_prediction_dataset.csv"
    df = pd.read_csv(file_path)
    return df

df = load_data()

# Encode categorical features
label_encoders = {}
for col in ["gender", "smoking_history"]:
    le = LabelEncoder()
    df[col] = le.fit_transform(df[col])
    label_encoders[col] = le

# Convert binary features (0,1) to "Yes" and "No" for display
binary_columns = ["hypertension", "heart_disease", "diabetes"]
df_display = df.copy()  # Keep a copy for display
for col in binary_columns:
    df_display[col] = df_display[col].map({0: "No", 1: "Yes"})

# Splitting dataset
X = df.drop(columns=["diabetes"])
y = df["diabetes"]  # Keep original 0/1 format

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Standardizing data
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

# Train Random Forest model
rf = RandomForestClassifier(n_estimators=100, random_state=42)
rf.fit(X_train_scaled, y_train)

# Tabs
tab1, tab2, tab3 = st.tabs(["📄 Dataset Preview", "📈 Model Performance", "🩺 Prediction"])

# 1️⃣ **Tab 1: Dataset Preview**
with tab1:
    st.subheader("📄 Complete Dataset Preview")
    st.write(df_display)  # Show dataset with Yes/No for better readability

    st.subheader("📊 Correlation Heatmap")
    plt.figure(figsize=(10,6))
    sns.heatmap(df.corr(), annot=True, cmap="coolwarm", fmt=".2f")
    st.pyplot(plt)

# 2️⃣ **Tab 2: Model Performance**
with tab2:
    st.subheader("📈 Model Performance")
    
    # Evaluate model
    y_pred = rf.predict(X_test_scaled)
    accuracy = accuracy_score(y_test, y_pred)
    st.write(f"### ⚡ Random Forest Accuracy: **{accuracy:.2f}**")

    # Confusion Matrix
    st.write("### 📊 Confusion Matrix")
    cm = confusion_matrix(y_test, y_pred)
    plt.figure(figsize=(5,4))
    sns.heatmap(cm, annot=True, fmt="d", cmap="Blues", xticklabels=["No Diabetes", "Diabetes"], yticklabels=["No Diabetes", "Diabetes"])
    plt.xlabel("Predicted")
    plt.ylabel("Actual")
    st.pyplot(plt)

    # ROC Curve
    st.write("### 📉 ROC Curve")
    fpr, tpr, _ = roc_curve(y_test, rf.predict_proba(X_test_scaled)[:,1])
    roc_auc = auc(fpr, tpr)
    plt.figure(figsize=(6,4))
    plt.plot(fpr, tpr, color='darkorange', lw=2, label='ROC curve (area = {:.2f})'.format(roc_auc))
    plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')
    plt.xlabel("False Positive Rate")
    plt.ylabel("True Positive Rate")
    plt.title("Receiver Operating Characteristic (ROC) Curve")
    plt.legend(loc="lower right")
    st.pyplot(plt)

# 3️⃣ **Tab 3: Prediction**
with tab3:
    st.subheader("🩺 Make a Prediction")

    # User inputs
    user_name = st.text_input("Patient Name", value="John Doe")
    user_gender = st.selectbox("Gender", label_encoders["gender"].classes_, key="gender_input")
    user_smoking = st.selectbox("Smoking History", label_encoders["smoking_history"].classes_, key="smoking_input")

    # Convert categorical inputs using label encoders
    user_gender_encoded = label_encoders["gender"].transform([user_gender])[0]
    user_smoking_encoded = label_encoders["smoking_history"].transform([user_smoking])[0]

    # User inputs numerical features
    user_data = [user_gender_encoded, user_smoking_encoded]
    for col in ["age", "bmi", "HbA1c_level", "blood_glucose_level"]:
        user_data.append(st.number_input(f"Enter {col}", float(df[col].min()), float(df[col].max()), float(df[col].mean())))

    # User inputs binary features
    user_binary_data = {}
    for col in ["hypertension", "heart_disease"]:
        user_binary_data[col] = st.radio(f"{col.replace('_', ' ').title()} (Yes/No)", ["No", "Yes"])

    # Convert "Yes"/"No" to numerical (0 or 1) before prediction
    for col in ["hypertension", "heart_disease"]:
        user_data.append(1 if user_binary_data[col] == "Yes" else 0)

    # Convert input into array
    user_data = np.array([user_data]).reshape(1, -1)

    # Predict button
    if st.button("🔮 Predict"):
        user_data_scaled = scaler.transform(user_data)
        
        # Prediction
        prediction = rf.predict(user_data_scaled)
        probability = rf.predict_proba(user_data_scaled)[:, 1][0]
        
        # Display result with patient name
        st.subheader(f"🤖 Prediction for {user_name}")
        if prediction[0] == 1:
            st.error(f"🚨 **{user_name} is likely to have diabetes.** (Probability: {probability:.2f})")
        else:
            st.success(f"✅ **{user_name} is not likely to have diabetes.** (Probability: {probability:.2f})")