app.py

import streamlit as st
import pandas as pd
import numpy as np
import pickle
import base64
import seaborn as sns
import matplotlib.pyplot as plt

import symbol

st.write("""
# Detection Fraud Credit Card

Kartu kredit adalah sebuah alat pembayaran menggunakan kartu yang berfungsi sebagai pengganti uang tunai.
""")

url_dataset = f'<a href="card.csv">Download Dataset CSV File</a>'
st.markdown(url_dataset, unsafe_allow_html=True)

def user_input_features() :
    distance_from_home = st.sidebar.slider('distance_from_home', 0.004874, 10632.723672)
    distance_from_last_transaction = st.sidebar.slider('distance_from_last_transaction', 0.000118, 11851.104565)
    ratio_to_median_purchase_price = st.sidebar.slider('ratio_to_median_purchase_price', 0.004399, 267.802942)
    repeat_retailer = st.sidebar.slider('repeat_retailer', 0.0, 1.0)
    used_chip = st.sidebar.slider('used_chip', 0.0, 1.0)
    used_pin_number  = st.sidebar.slider('used_pin_number ', 0.0, 1.0)
    online_order  = st.sidebar.slider('online_order ', 0.0, 1.0)

    data = {
    'distance_from_home':[distance_from_home],
    'distance_from_last_transaction':[distance_from_last_transaction],
    'ratio_to_median_purchase_price':[ratio_to_median_purchase_price],
    'repeat_retailer':[repeat_retailer],
    'used_pin_number':[used_pin_number],
    'online_order':[online_order],
    'used_chip':[used_chip]
    }

    features = pd.DataFrame(data)
    return features

input_df = user_input_features()

card_raw = pd.read_csv('card.csv')
card_raw.fillna(0, inplace=True)
card = card_raw.drop(columns=['fraud'])
df = pd.concat([input_df, card],axis=0)

df = df[:1] # Selects only the first row (the user input data)
df.fillna(0, inplace=True)

features = ['distance_from_home', 'distance_from_last_transaction',
       'ratio_to_median_purchase_price', 'repeat_retailer', 'used_chip',
       'used_pin_number', 'online_order']

df = df[features]

st.subheader('User Input features')
st.write(df)
load_clf = pickle.load(open('card_clf.pkl', 'rb'))
detection = load_clf.predict(df)
if(detection > 0) :
    detection = 1
detection_proba = load_clf.predict_proba(df)
knee_labels = np.array(['Normal','Penipuan'])
st.subheader('Detection')
st.write(knee_labels[detection])
st.subheader('Detection Probability')
df_prob = pd.DataFrame(data=detection_proba, index=['Probability'], columns=knee_labels)
st.write(df_prob)