app.py

import gradio as gr
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans
from sklearn.tree import DecisionTreeClassifier

# -----------------------------
# Dummy Dataset (Replace with your actual model if available)
# -----------------------------
np.random.seed(42)

data = []
for _ in range(200):
    data.append([
        np.random.uniform(0.0, 1.0),   # GreenPurchaseRatio
        np.random.uniform(1.0, 7.0),   # CarbonFootprintScore
        np.random.uniform(0.0, 1.0),   # RecyclingRate
        np.random.uniform(0, 30),      # EcoPremiumWillingness
        np.random.randint(1, 40),      # PurchaseFrequency
        np.random.randint(0, 100)      # SustainabilityScore
    ])

columns = [
    "GreenPurchaseRatio",
    "CarbonFootprintScore",
    "RecyclingRate",
    "EcoPremiumWillingness",
    "PurchaseFrequency",
    "SustainabilityScore"
]

df_dummy = pd.DataFrame(data, columns=columns)

# -----------------------------
# Train Models
# -----------------------------
scaler = StandardScaler()
X_scaled = scaler.fit_transform(df_dummy)

kmeans = KMeans(n_clusters=4, random_state=42, n_init=10)
clusters = kmeans.fit_predict(X_scaled)

df_dummy["Cluster"] = clusters

tree = DecisionTreeClassifier(max_depth=5, random_state=42)
tree.fit(X_scaled, clusters)

# -----------------------------
# Cluster Names
# -----------------------------
cluster_names = {
    0: "Price Driven",
    1: "Eco Indifferent",
    2: "Green Champion",
    3: "Eco Curious"
}

# -----------------------------
# Marketing Recommendations
# -----------------------------
recommendations = {
    "Green Champion": """
Premium Eco Products

Sustainability Rewards

Carbon Offset Programs

Exclusive Green Membership
""",

    "Eco Curious": """
Sustainability Awareness Campaigns

First-Time Green Discounts

Eco Product Recommendations

Loyalty Incentives
""",

    "Price Driven": """
Cost Saving Offers

Bundle Discounts

Budget Friendly Eco Products

Cashback Campaigns
""",

    "Eco Indifferent": """
Convenience-Based Promotions

Product Quality Campaigns

Personalized Deals

Retention Strategies
"""
}

# -----------------------------
# Prediction Function
# -----------------------------
def predict_segment(
    gpr,
    carbon,
    recycle,
    premium,
    freq,
    score
):

    sample = pd.DataFrame([[
        gpr,
        carbon,
        recycle,
        premium,
        freq,
        score
    ]], columns=columns)

    sample_scaled = scaler.transform(sample)

    pred = tree.predict(sample_scaled)[0]

    segment = cluster_names[pred]

    strategy = recommendations[segment]

    return segment, strategy


# -----------------------------
# Gradio UI
# -----------------------------
with gr.Blocks(theme=gr.themes.Soft()) as demo:

    gr.Markdown("""
# 🌿 GreenLeaf Retail Analytics Dashboard

### AI-Powered Customer Segmentation System

This system uses:

- K-Means Clustering
- Decision Tree Classification

to identify customer segments and recommend targeted marketing strategies.
""")

    with gr.Row():

        with gr.Column():

            gr.Markdown("## Customer Inputs")

            gpr = gr.Slider(
                0,
                1,
                value=0.5,
                label="Green Purchase Ratio"
            )

            carbon = gr.Slider(
                1,
                7,
                value=3,
                label="Carbon Footprint Score"
            )

            recycle = gr.Slider(
                0,
                1,
                value=0.5,
                label="Recycling Engagement Rate"
            )

            premium = gr.Slider(
                0,
                30,
                value=10,
                label="Eco Premium Willingness (%)"
            )

            freq = gr.Slider(
                1,
                40,
                value=15,
                label="Purchase Frequency"
            )

            score = gr.Slider(
                0,
                100,
                value=50,
                label="Sustainability Score"
            )

            predict_btn = gr.Button(
                "Predict Customer Segment",
                variant="primary"
            )

        with gr.Column():

            gr.Markdown("## Prediction Result")

            segment_output = gr.Textbox(
                label="Customer Segment",
                lines=2
            )

            recommendation_output = gr.Textbox(
                label="Recommended Marketing Strategy",
                lines=10
            )

    predict_btn.click(
        fn=predict_segment,
        inputs=[
            gpr,
            carbon,
            recycle,
            premium,
            freq,
            score
        ],
        outputs=[
            segment_output,
            recommendation_output
        ]
    )

    with gr.Accordion("Model Information", open=False):

        gr.Markdown("""
### Project Details

#### Dataset Features
- Green Purchase Ratio
- Carbon Footprint Score
- Recycling Engagement Rate
- Eco Premium Willingness
- Purchase Frequency
- Sustainability Score

#### Algorithms Used
- K-Means Clustering (K = 4)
- Decision Tree Classifier

#### Customer Segments
- Green Champion
- Eco Curious
- Price Driven
- Eco Indifferent

#### Business Objective
Customer Segmentation for Green Retail Marketing.
""")

demo.launch()