models.py

"""
Graded Challenge 5
Nama: Qothrunnadaa Alyaa
Batch: HCK-009

File ini digunakan untuk menjalankan form prediksi default nasabah kartu kredit
"""

import streamlit as st
import pandas as pd
import pickle

# Membuat function untuk dipanggil di app.py
def run():
    st.title('Credit Card Default Predictor')

    # Mendefinisikan pilihan jenjang pendidikan terakhir
    edu_lvl = ['High School', 'Undergraduate', 'Graduate School', 'Others']

    # Mendefinisikan pilihan status pembayaran tagihan
    pay_option_map = {
        '-2: Unused': -2,
        '-1: Paid in full': -1,
        '0: Revolving credit': 0,
        '1: One month late payment': 1,
        '2: Two months late payment': 2,
        '3: Three months late payment': 3,
        '4: Four months late payment': 4,
        '5: Five months late payment': 5,
        '6: Six months late payment': 6,
        '7: Seven months late payment': 7,
        '8: Eight months late payment': 8,
        '9: Nine months or more late payment': 9
    }
    pay_option = list(pay_option_map.keys())
    pay_status = {}
    
    # Mendefinisikan nominal tagihan yang dibayarkan oleh nasabah
    paid_amt = {}
    
    # List bulan tagihan
    months = ['September', 'August', 'July', 'June', 'May', 'April']

    # Memasukkan data inference dengan form prediksi

    # Limit balance
    limit_balance = st.number_input(label='Input your credit card limit balance here', min_value=0.0)

    # Education level
    education_level = st.selectbox(label='Education Level:', options=edu_lvl)
    if education_level == 'Graduate School':
        education = 1
    elif education_level == 'High School':
        education = 2
    elif education_level == 'Undergraduate':
        education = 3
    else:
        education = 0

    # Status pembayaran (pay) dan tagihan yang dibayarkan (pay_amt)
    for month in months:
        chosen_month = st.selectbox(label=f'Payment status in {month}', options=pay_option)
        pay_status[month] = pay_option_map[chosen_month]
        paid_amt[month] = st.number_input(label=f'Paid amount in {month}', min_value=0.0)

    # Membuat data inference dari data yang dimasukkin
    data_inf = pd.DataFrame({
        'limit_balance': limit_balance,
        'education_level': education,
        'pay_9': pay_status['September'],
        'pay_8': pay_status['August'],
        'pay_7': pay_status['July'],
        'pay_6': pay_status['June'],
        'pay_5': pay_status['May'],
        'pay_4': pay_status['April'],
        'pay_amt_9': paid_amt['September'],
        'pay_amt_8': paid_amt['August'],
        'pay_amt_7': paid_amt['July'],
        'pay_amt_6': paid_amt['June'],
        'pay_amt_5': paid_amt["May"],
        'pay_amt_4': paid_amt['April']
    }, index=[0])
    
    # Menampilkan data 
    st.header("Client's Credit Overview")
    st.table(data_inf)

    # Memprediksi apakah nasabah akan mengalami default atau tidak
    if st.button(label='Predict'):
        with open('logreg.pkl', 'rb') as logreg:
            model = pickle.load(logreg)
            
        y_pred_inf = model.predict(data_inf)
        
        if y_pred_inf == 0:
            st.write('Nasabah tidak gagal bayar tagihan bulan depan')
        else:
            st.write('Nasabah akan GAGAL BAYAR bulan depan')