app.py

import streamlit as st
import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestRegressor
import joblib

st.title('Restaurant Revenue Predictor')

# Create input form
st.write('Enter restaurant details:')

# City selection
city = st.selectbox('City', ['Istanbul', 'Ankara', 'Izmir', 'Other Cities'])

# Adding description for Type
st.write("""
**Type**: Type of the restaurant
- FC: Food Court
- IL: Inline
- DT: Drive Thru
- MB: Mobile
""")
type = st.selectbox('Type', ['FC', 'IL', 'DT', 'MB'])

# Create a simple model for demonstration
@st.cache_resource
def create_model():
    model = RandomForestRegressor(
        n_estimators=100,
        max_depth=10,
        random_state=42
    )
    # Create some sample training data
    X_train = pd.DataFrame({
        'City Group_Big Cities': [1, 0, 1, 0],
        'City Group_Other': [0, 1, 0, 1],
        'Type_DT': [1, 0, 0, 0],
        'Type_FC': [0, 1, 0, 0],
        'Type_IL': [0, 0, 1, 0],
        'Type_MB': [0, 0, 0, 1],
        'days': [1000, 800, 600, 400]
    })
    y_train = np.array([1500000, 1000000, 800000, 500000])
    model.fit(X_train, y_train)
    return model

if st.button('Predict Revenue'):
    # Map city to City Group
    city_group = 'Big Cities' if city in ['Istanbul', 'Ankara', 'Izmir'] else 'Other'
    
    # Create input dataframe
    input_data = pd.DataFrame({
        'City Group_Big Cities': [1 if city_group == 'Big Cities' else 0],
        'City Group_Other': [1 if city_group == 'Other' else 0],
        'Type_DT': [1 if type == 'DT' else 0],
        'Type_FC': [1 if type == 'FC' else 0],
        'Type_IL': [1 if type == 'IL' else 0],
        'Type_MB': [1 if type == 'MB' else 0],
        'days': [500]  # default value
    })
    
    try:
        # Get model
        model = create_model()
        
        # Make prediction
        prediction = model.predict(input_data)[0]
        
        # Format prediction
        formatted_prediction = f"${prediction:,.2f}"
        
        # Display prediction with additional context
        st.success(f'Predicted Revenue: {formatted_prediction}')
        
        # Add some context about the prediction
        st.info("""
        Note: This is a simplified model for demonstration purposes. 
        The prediction is based on limited training data and should be used as a rough estimate only.
        """)
        
    except Exception as e:
        st.error(f"Error making prediction: {str(e)}")