src/streamlit_app.py

import streamlit as st
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier, plot_tree
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report
from sklearn.preprocessing import LabelEncoder

# Konfigurasi Halaman
st.set_page_config(page_title="Bank Marketing DT App", layout="wide")

st.title("🌳 Decision Tree Classifier: Bank Marketing")
st.markdown("""
Aplikasi ini diadaptasi dari notebook pertemuan EDA & Decision Tree. 
Silakan unggah dataset **bank-full.csv** untuk memulai analisis.
""")

# Sidebar untuk Input Data
st.sidebar.header("📂 Menu Upload")
uploaded_file = st.sidebar.file_uploader("Pilih file CSV", type=["csv"])

if uploaded_file is not None:
    # Membaca data (Gunakan sep=';' karena dataset bank marketing biasanya menggunakan titik koma)
    try:
        df = pd.read_csv(uploaded_file, sep=';')
    except Exception as e:
        st.error(f"Error membaca file: {e}")
        st.stop()

    # Navigasi di Sidebar
    menu = st.sidebar.radio("Navigasi", ["Eksplorasi Data (EDA)", "Modeling & Evaluasi"])

    if menu == "Eksplorasi Data (EDA)":
        st.subheader("🔍 Statistik Deskriptif & Preview")
        
        col1, col2 = st.columns([2, 1])
        with col1:
            st.write("5 Data Teratas:")
            st.dataframe(df.head())
        with col2:
            st.write("Informasi Dataset:")
            st.write(f"Baris: {df.shape[0]}")
            st.write(f"Kolom: {df.shape[1]}")
            st.write(df.dtypes)

        st.divider()
        
        st.subheader("📊 Visualisasi Distribusi Target")
        fig, ax = plt.subplots(figsize=(8, 4))
        sns.countplot(data=df, x='y', palette='viridis', ax=ax)
        st.pyplot(fig)

    elif menu == "Modeling & Evaluasi":
        st.subheader("⚙️ Training Decision Tree")
        
        # Preprocessing Sederhana sesuai Notebook
        df_model = df.copy()
        
        # Encoding categorical variables
        le = LabelEncoder()
        for col in df_model.select_dtypes(include=['object']).columns:
            df_model[col] = le.fit_transform(df_model[col])
        
        # Split Fitur dan Target
        X = df_model.drop('y', axis=1)
        y = df_model['y']
        
        # Pengaturan Hyperparameter di Sidebar
        st.sidebar.subheader("Hyperparameters")
        test_size = st.sidebar.slider("Test Size (%)", 10, 50, 30)
        max_depth = st.sidebar.slider("Max Depth Tree", 1, 20, 5)
        
        X_train, X_test, y_train, y_test = train_test_split(
            X, y, test_size=test_size/100, random_state=42
        )

        if st.button("🚀 Train Model"):
            model = DecisionTreeClassifier(max_depth=max_depth, random_state=42)
            model.fit(X_train, y_train)
            y_pred = model.predict(X_test)

            # Hasil Metrik
            acc = accuracy_score(y_test, y_pred)
            
            col_m1, col_m2 = st.columns(2)
            col_m1.metric("Accuracy Score", f"{acc:.2%}")
            
            # Confusion Matrix
            st.write("### Confusion Matrix")
            cm = confusion_matrix(y_test, y_pred)
            fig_cm, ax_cm = plt.subplots()
            sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', ax=ax_cm)
            plt.xlabel('Predicted')
            plt.ylabel('Actual')
            st.pyplot(fig_cm)

            # Visualisasi Pohon (Terbatas kedalaman agar tidak berat)
            st.write("### Pohon Keputusan (Visualisasi)")
            fig_tree, ax_tree = plt.subplots(figsize=(20, 10))
            plot_tree(model, feature_names=X.columns, class_names=['No', 'Yes'], 
                      filled=True, max_depth=3, ax=ax_tree, fontsize=10)
            st.pyplot(fig_tree)
            
            st.write("### Classification Report")
            st.text(classification_report(y_test, y_pred))

else:
    st.warning("👈 Silakan unggah dataset di sidebar untuk memproses data.")
    st.info("Catatan: Gunakan dataset 'bank-full.csv' dari UCI Machine Learning Repository.")