app.py

import gradio as gr
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
import warnings

# Ignore all warnings
warnings.filterwarnings("ignore")

# Load Data
df_lending_data = pd.read_csv('Resources/lending_data.csv')

# Prepare Features and Labels
y = df_lending_data['loan_status']
X = df_lending_data.drop(columns=['loan_status'])

# Split Data
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1)

# Train Logistic Regression Model
model = LogisticRegression(max_iter=200, random_state=1)
model.fit(X_train, y_train)

# Gradio Function for Prediction
def predict_loan_status(loan_size, interest_rate, borrower_income, debt_to_income, num_of_accounts, derogatory_marks, total_debt):
    input_data = pd.DataFrame({
        'loan_size': [loan_size],
        'interest_rate': [interest_rate],
        'borrower_income': [borrower_income],
        'debt_to_income': [debt_to_income],
        'num_of_accounts': [num_of_accounts],
        'derogatory_marks': [derogatory_marks],
        'total_debt': [total_debt]
    })
    prediction = model.predict(input_data)
    return "Healthy Loan (0)" if prediction[0] == 0 else "High-Risk Loan (1)"

# Create the Gradio interface
interface = gr.Interface(
    fn=predict_loan_status,
    inputs=[
        gr.Number(label="Loan Size"),
        gr.Number(label="Interest Rate"),
        gr.Number(label="Borrower Income"),
        gr.Number(label="Debt-to-Income Ratio"),
        gr.Number(label="Number of Accounts"),
        gr.Number(label="Derogatory Marks"),
        gr.Number(label="Total Debt"),
    ],
    outputs="text",
    title="Loan Status Prediction",
    description="Input loan details to predict whether the loan is healthy or high-risk. "
                "Explore the project on GitHub [here](https://github.com/thaychansy/credit-risk-classification) "
                "and try out other models [here](https://huggingface.co/tchans123)."
)

# Launch Gradio interface
interface.launch(share=True)