app.py

import streamlit as st
import pandas as pd
import numpy as np
import joblib

from sklearn.preprocessing import LabelEncoder
from sklearn.impute import SimpleImputer


# =========================================================
# PAGE CONFIG
# =========================================================

st.set_page_config(
    page_title="Finora",
    page_icon="💳",
    layout="wide",
    initial_sidebar_state="expanded"
)


# =========================================================
# CUSTOM CSS
# =========================================================

st.markdown("""
<style>

@import url('https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700;800;900&display=swap');

html, body, [class*="css"] {
    font-family: 'Inter', sans-serif;
}

.main {
    background-color: #F8FAFC;
}

/* Sidebar */

section[data-testid="stSidebar"] {
    background: linear-gradient(180deg, #0F172A 0%, #111827 100%);
    border-right: 1px solid #1E293B;
}

section[data-testid="stSidebar"] * {
    color: white !important;
}

/* Hero */

.hero {
    background: linear-gradient(135deg, #0F172A 0%, #1E293B 100%);
    padding: 3rem;
    border-radius: 24px;
    color: white;
    margin-bottom: 2rem;
    box-shadow: 0px 10px 30px rgba(0,0,0,0.15);
}

.hero-title {
    font-size: 3rem;
    font-weight: 800;
}

.hero-subtitle {
    font-size: 1.1rem;
    opacity: 0.85;
    margin-top: 1rem;
}

/* Cards */

.metric-card {
    background: white;
    padding: 1.5rem;
    border-radius: 20px;
    border: 1px solid #E2E8F0;
    box-shadow: 0px 4px 20px rgba(15,23,42,0.06);
    transition: 0.3s ease;
}

.metric-card:hover {
    transform: translateY(-4px);
    box-shadow: 0px 12px 30px rgba(15,23,42,0.12);
}

.metric-title {
    color: #64748B;
    font-size: 0.9rem;
    font-weight: 600;
}

.metric-value {
    color: #0F172A;
    font-size: 2rem;
    font-weight: 800;
}

/* Upload Box */

.upload-box {
    background: white;
    padding: 2rem;
    border-radius: 24px;
    border: 2px dashed #CBD5E1;
    box-shadow: 0px 4px 20px rgba(15,23,42,0.06);
}

/* Buttons */

.stButton>button {
    background: linear-gradient(135deg, #D4AF37 0%, #FBBF24 100%);
    color: #0F172A;
    border: none;
    border-radius: 12px;
    font-weight: 700;
    padding: 0.8rem 1.5rem;
}

/* Dataframes */

[data-testid="stDataFrame"] {
    border-radius: 20px;
    overflow: hidden;
    border: 1px solid #E2E8F0;
}

/* Footer */

.footer {
    text-align: center;
    padding: 2rem;
    color: #64748B;
    font-size: 0.9rem;
}

</style>
""", unsafe_allow_html=True)


# =========================================================
# LOAD MODEL
# =========================================================

model = joblib.load('outputs/best_model.pkl')


# =========================================================
# SIDEBAR
# =========================================================

with st.sidebar:

    st.markdown("## 💳 Finora")

    st.markdown("---")

    page = st.radio(
        "Navigation",
        [
            "🏠 Dashboard",
            "📁 Upload & Predict",
            "📊 Analytics",
            "❓ Platform Info"
        ]
    )

    st.markdown("---")

    st.success("System Active")

    st.markdown("""
    ### Platform Status

    ✔ AI Prediction Engine Online  
    ✔ Risk Monitoring Active  
    ✔ Loan Analytics Enabled  
    """)


# =========================================================
# HERO SECTION
# =========================================================

st.markdown("""
<div class="hero">

<div class="hero-title">
Finora — AI-Powered Microfinance Intelligence Platform
</div>

<div class="hero-subtitle">
Enterprise-grade fintech analytics platform for telecom microfinance
repayment prediction and customer risk intelligence.
</div>

</div>
""", unsafe_allow_html=True)


# =========================================================
# KPI SECTION
# =========================================================

c1, c2, c3, c4 = st.columns(4)

with c1:
    st.markdown("""
    <div class="metric-card">
        <div class="metric-title">Model Accuracy</div>
        <div class="metric-value">92%</div>
    </div>
    """, unsafe_allow_html=True)

with c2:
    st.markdown("""
    <div class="metric-card">
        <div class="metric-title">Loans Processed</div>
        <div class="metric-value">209K+</div>
    </div>
    """, unsafe_allow_html=True)

with c3:
    st.markdown("""
    <div class="metric-card">
        <div class="metric-title">Prediction Engine</div>
        <div class="metric-value">ACTIVE</div>
    </div>
    """, unsafe_allow_html=True)

with c4:
    st.markdown("""
    <div class="metric-card">
        <div class="metric-title">Risk Monitoring</div>
        <div class="metric-value">LIVE</div>
    </div>
    """, unsafe_allow_html=True)


st.markdown("<br>", unsafe_allow_html=True)


# =========================================================
# DASHBOARD PAGE
# =========================================================

if page == "🏠 Dashboard":

    st.markdown("## 🏠 Executive Dashboard")

    d1, d2, d3 = st.columns(3)

    with d1:
        st.metric("Prediction Accuracy", "92%")

    with d2:
        st.metric("Non Defaulter Rate", "87.5%")

    with d3:
        st.metric("Defaulter Rate", "12.5%")

    st.info(
        "AI-driven telecom behavioral analytics for "
        "microfinance repayment prediction."
    )


# =========================================================
# UPLOAD PAGE
# =========================================================

elif page == "📁 Upload & Predict":

    st.markdown("""
    <div class="upload-box">

    <h3 style='color:#0F172A;'>
    Upload Customer Dataset
    </h3>

    <p style='color:#64748B;'>
    Upload telecom microfinance customer data to generate
    AI-powered repayment predictions.
    </p>

    </div>
    """, unsafe_allow_html=True)

    uploaded_file = st.file_uploader(
        "Upload CSV File",
        type=['csv']
    )

    if uploaded_file is not None:

        with st.spinner("Analyzing repayment behavior..."):

            data = pd.read_csv(uploaded_file)

            original_data = data.copy()

            # Drop unwanted columns

            drop_cols = ['msisdn']

            for col in drop_cols:

                if col in data.columns:

                    data.drop(col, axis=1, inplace=True)

            # Remove target if exists

            if 'label' in data.columns:

                data.drop('label', axis=1, inplace=True)

            # Handle missing values

            num_cols = data.select_dtypes(
                include=np.number
            ).columns

            imputer = SimpleImputer(strategy='median')

            data[num_cols] = imputer.fit_transform(
                data[num_cols]
            )

            # Encode categoricals

            cat_cols = data.select_dtypes(
                include='object'
            ).columns

            encoder = LabelEncoder()

            for col in cat_cols:

                data[col] = encoder.fit_transform(
                    data[col].astype(str)
                )

            # Predictions

            prediction_probability = model.predict_proba(data)[:, 1]

            prediction_label = model.predict(data)

            prediction_status = np.where(
                prediction_label == 1,
                'Non Defaulter',
                'Defaulter'
            )

            # Output

            original_data['Repayment_Probability'] = (
                prediction_probability
            )

            original_data['Prediction_Label'] = (
                prediction_label
            )

            original_data['Prediction_Status'] = (
                prediction_status
            )

        st.success("Prediction completed successfully!")

        st.markdown("## 📊 Prediction Results")

        st.dataframe(
            original_data.head(50),
            use_container_width=True
        )

        # Metrics

        success_rate = (
            (prediction_label == 1).mean() * 100
        )

        risk_rate = (
            (prediction_label == 0).mean() * 100
        )

        avg_probability = (
            prediction_probability.mean() * 100
        )

        m1, m2, m3 = st.columns(3)

        with m1:
            st.metric(
                "Non Defaulters",
                f"{success_rate:.2f}%"
            )

        with m2:
            st.metric(
                "Defaulters",
                f"{risk_rate:.2f}%"
            )

        with m3:
            st.metric(
                "Avg Repayment Score",
                f"{avg_probability:.2f}%"
            )

        csv = original_data.to_csv(index=False)

        st.download_button(
            label="⬇ Download Prediction Report",
            data=csv,
            file_name="prediction_results.csv",
            mime="text/csv"
        )


# =========================================================
# ANALYTICS PAGE
# =========================================================

elif page == "📊 Analytics":

    st.markdown("## 📊 Portfolio Analytics")

    a1, a2 = st.columns(2)

    with a1:
        st.info(
            "AI-driven repayment prediction engine monitoring "
            "telecom microfinance customer risk behavior."
        )

    with a2:
        st.success(
            "Real-time intelligent classification system for "
            "loan repayment probability forecasting."
        )

    st.markdown("### Risk Intelligence")

    st.progress(92)

    st.markdown("""
    ✔ High-performing ML prediction engine

    ✔ Intelligent customer risk segmentation

    ✔ Automated telecom repayment analytics

    ✔ AI-powered default prediction
    """)


# =========================================================
# PLATFORM INFO
# =========================================================

elif page == "❓ Platform Info":

    st.markdown("## ❓ Platform Information")

    with st.expander("What does this platform do?"):

        st.write("""
        This AI-powered fintech platform predicts whether
        a telecom microfinance customer is likely to repay
        a loan within the repayment period.
        """)

    with st.expander("Prediction Labels"):

        st.write("""
        - **1 → Non Defaulter**
        - **0 → Defaulter**
        """)

    with st.expander("Technology Stack"):

        st.write("""
        - Streamlit
        - Scikit-Learn
        - Random Forest
        - CatBoost
        - Machine Learning Analytics
        """)


# =========================================================
# FOOTER
# =========================================================

st.markdown("""
<div class="footer">

© 2026 Finora • Enterprise Microfinance Intelligence Platform

</div>
""", unsafe_allow_html=True)