app.py

import pandas as pd
import streamlit as st
import numpy as np
import pickle
import catboost
from sklearn.impute import SimpleImputer



# Load the saved model and unique values:
with open("model_and_key_components.pkl", "rb") as f:
    components = pickle.load(f)

# Extract the individual components
dt_model = components["model"]
unique_values = components["unique_values"]


st.image("https://i.ytimg.com/vi/WULwst0vW8g/maxresdefault.jpg")
st.title("Income Prediction App")

# Sidebar with input field descriptions
st.sidebar.header("Description of the Required Input Fields")
st.sidebar.markdown("**Age**: Enter the age of the individual (e.g., 25, 42, 57).")
st.sidebar.markdown("**Gender**: Select the gender of the individual (e.g., Male, Female).")
st.sidebar.markdown("**Education**: Choose the highest education level of the individual (e.g., Bachelors Degree, High School Graduate, Masters Degree).")
st.sidebar.markdown("**Worker Class**: Select the class of worker for the individual (e.g., Private, Government, Self-employed).")
st.sidebar.markdown("**Marital Status**: Choose the marital status of the individual (e.g., Married, Never married, Divorced).")
st.sidebar.markdown("**Race**: Select the race of the individual (e.g., White, Black, Asian-Pac-Islander).")
st.sidebar.markdown("**Hispanic Origin**: Choose the Hispanic origin of the individual (e.g., Mexican, Puerto Rican, Cuban).")
st.sidebar.markdown("**Full/Part-Time Employment**: Select the employment status as full-time or part-time (e.g., Full-time schedules, Part-time schedules).")
st.sidebar.markdown("**Wage Per Hour**: Enter the wage per hour of the individual (numeric value, e.g., 20.50).")
st.sidebar.markdown("**Weeks Worked Per Year**: Specify the number of weeks the individual worked in a year (numeric value, e.g., 45).")
st.sidebar.markdown("**Industry Code**: Choose the category code of the industry where the individual works (e.g., Category 1, Category 2).")
st.sidebar.markdown("**Major Industry Code**: Select the major industry code of the individual's work (e.g., Industry A, Industry B).")
st.sidebar.markdown("**Occupation Code**: Choose the category code of the occupation of the individual (e.g., Category X, Category Y).")
st.sidebar.markdown("**Major Occupation Code**: Select the major occupation code of the individual (e.g., Occupation 1, Occupation 2).")
st.sidebar.markdown("**Total Employed**: Specify the number of persons worked for the employer (numeric value, e.g., 3, 5).")
st.sidebar.markdown("**Household Stat**: Choose the detailed household and family status of the individual (e.g., Single, Married-civilian spouse present).")
st.sidebar.markdown("**Household Summary**: Select the detailed household summary (e.g., Child under 18 never married, Spouse of householder).")
st.sidebar.markdown("**Veteran Benefits**: Choose whether the individual receives veteran benefits (Yes or No).")
st.sidebar.markdown("**Tax Filer Status**: Select the tax filer status of the individual (e.g., Single, Joint both 65+).")
st.sidebar.markdown("**Gains**: Specify any gains the individual has (numeric value, e.g., 1500.0).")
st.sidebar.markdown("**Losses**: Specify any losses the individual has (numeric value, e.g., 300.0).")
st.sidebar.markdown("**Dividends from Stocks**: Specify any dividends from stocks for the individual (numeric value, e.g., 120.5).")
st.sidebar.markdown("**Citizenship**: Select the citizenship status of the individual (e.g., Native, Foreign Born- Not a citizen of U S).")
st.sidebar.markdown("**Year of Migration**: Enter the year of migration for the individual (numeric value, e.g., 2005).")
st.sidebar.markdown("**Country of Birth**: Choose the individual's birth country (e.g., United-States, Other).")
st.sidebar.markdown("**Importance of Record**: Enter the weight of the instance (numeric value, e.g., 0.9).")

# Create input fields for user input
col1, col2, col3 = st.columns(3)

with col1:
    age = st.number_input("Age", min_value=0)
    gender = st.selectbox("Gender", ["Male", "Female"])
    education = st.selectbox("Education", unique_values['education'])
    worker_class = st.selectbox("Class of Worker", unique_values['worker_class'])
    marital_status = st.selectbox("Marital Status", unique_values['marital_status'])
    race = st.selectbox("Race", unique_values['race'])
    is_hispanic = st.selectbox("Hispanic Origin", unique_values['is_hispanic'])
    employment_commitment = st.selectbox("Full/Part-Time Employment", unique_values['employment_commitment'])
    wage_per_hour = st.number_input("Wage Per Hour", min_value=0)

with col2:
    working_week_per_year = st.number_input("Weeks Worked Per Year", min_value=0)
    industry_code = st.selectbox("Category Code of Industry", unique_values['industry_code'])
    industry_code_main = st.selectbox("Major Industry Code", unique_values['industry_code_main'])
    occupation_code = st.selectbox("Category Code of Occupation", unique_values['occupation_code'])
    occupation_code_main = st.selectbox("Major Occupation Code", unique_values['occupation_code_main'])
    total_employed = st.number_input("Number of Persons Worked for Employer", min_value=0)
    household_stat = st.selectbox("Detailed Household and Family Status", unique_values['household_stat'])
    household_summary = st.selectbox("Detailed Household Summary", unique_values['household_summary'])
    vet_benefit = st.selectbox("Veteran Benefits", unique_values['vet_benefit'])

with col3:
    tax_status = st.selectbox("Tax Filer Status", unique_values['tax_status'])
    gains = st.number_input("Gains", min_value=0)
    losses = st.number_input("Losses", min_value=0)
    stocks_status = st.number_input("Dividends from Stocks", min_value=0)
    citizenship = st.selectbox("Citizenship", unique_values['citizenship'])
    mig_year = st.selectbox("Migration Year", unique_values['mig_year'])
    country_of_birth_own = st.selectbox("Country of Birth", unique_values['country_of_birth_own'])
    importance_of_record = st.number_input("Importance of Record", min_value=0.0)

# Button to trigger prediction
if st.button("Predict"):
    # Create a dictionary of user input
    user_input = {
        "age": age,
        "gender": gender,
        "education": education,
        "worker_class": worker_class,
        "marital_status": marital_status,
        "race": race,
        "is_hispanic": is_hispanic,
        "employment_commitment": employment_commitment,
        "wage_per_hour": wage_per_hour,
        "working_week_per_year": working_week_per_year,
        "industry_code": industry_code,
        "industry_code_main": industry_code_main,
        "occupation_code": occupation_code,
        "occupation_code_main": occupation_code_main,
        "total_employed": total_employed,
        "household_stat": household_stat,
        "household_summary": household_summary,
        "vet_benefit": vet_benefit,
        "tax_status": tax_status,
        "gains": gains,
        "losses": losses,
        "stocks_status": stocks_status,
        "citizenship": citizenship,
        "mig_year": mig_year,
        "country_of_birth_own": country_of_birth_own,
        "importance_of_record": importance_of_record
    }
    # Send a POST request to the FastAPI server
    response = requests.post("rasmodev-income-prediction-fastapi.hf.space/docs#/default/predict_income_predict__post", json=user_input)
    
    # Check if the request was successful
    if response.status_code == 200:
        prediction_data = response.json()
        
        # Display prediction result to the user
        st.subheader("Prediction Result")
    
        # Determine income prediction and format message
        if prediction_data['income_prediction'] == "Income over $50K":
            st.success("This individual is predicted to have an income of over $50K.")
        else:
            st.error("This individual is predicted to have an income of under $50K")
        
        # Display prediction probability
        st.subheader("Prediction Probability")
        probability = prediction_data['prediction_probability']
        st.write(f"The probability of the individual having an income over $50K is: {probability:.2f}")
    else:
        st.error("Error: Unable to get prediction")