app.py

import streamlit as st
import pandas as pd
import pickle

with open('gb_model_best.pkl', 'rb') as f:
    gb_model_best = pickle.load(f)

# Create a Streamlit app
st.title("Movie Revenue Predictor 🎥")
st.image("https://i0.wp.com/indyfilmlibrary.com/wp-content/uploads/2024/01/iStock-92043739-e1704448388494.jpg?fit=1329%2C578&ssl=1")
st.write("This data is trained from imdb with %70 r2 score. Learn which artibutes make a money machine film.")

# Create input fields for the user
popularity = st.slider("Popularity (0-100, 100 being Avatar):", min_value=1, max_value=100, value=50)
runtime = st.number_input("Runtime (in minutes):", min_value=1, step=1, value=120)
is_original_en = st.selectbox("Is the orginal language English?", ["Yes", "No"], index=0)

# Create a conditional input field for budget
budget_known = st.selectbox("Is the budget known?", ["Yes", "No"], index=0)
if budget_known == "Yes":
    budget_M = st.number_input("Budget (in millions of dollars):", min_value=1, step=1, value=150)
else:
    budget_M = 10

# Create a submit button
submitted = st.button("Predict Revenue")

# When the user submits the form, make a prediction
if submitted:
    popularity_value = popularity / 2
    # Create a DataFrame with the user's input data
    input_data = pd.DataFrame({
        "popularity": [popularity_value],
        "runtime": [runtime],
        "budget_known": [int(budget_known == "Yes")],
        "budget_M": [budget_M],
        "is_original_en": [int(is_original_en == "Yes")]
    })

    # Make a prediction using the trained model
    input_data = input_data[gb_model_best.feature_names_in_]
    prediction = gb_model_best.predict(input_data)[0]

    # Display the predicted revenue
    st.write(f"Predicted Box Office Revenue is:")
    st.title(f"${prediction:.2f} million.")