app.py

#Import Relevant Libraries
import streamlit as st
import pickle
import pandas as pd
from catboost import CatBoostClassifier

# Load the trained model and unique values from the pickle file
with open('model_and_key_components.pkl', 'rb') as file:
    saved_components = pickle.load(file)

model = saved_components['model']
unique_values = saved_components['unique_values']

# Page Title with Style
st.markdown(
    f"""
    <div style="text-align: center;">
        <h2 style="color: #3a2a71;">👨‍💼👩‍💼 Employee Attrition Prediction App</h2>
    </div>
    """,
    unsafe_allow_html=True
)

# Welcome Message with Style (Centered)
st.markdown(
    f"""
    <div style="text-align: center;">
        <p>Welcome to the Employee Attrition Prediction App! 🚀</p>
    </div>
    """,
    unsafe_allow_html=True
)

# Sepsis Information
st.markdown(
    """
    **Employee attrition** refers to the phenomenon of employees leaving their jobs for various reasons. It's crucial for organizations to predict attrition to retain valuable talent.
    """
)

# Link to Detailed Article on Employee Attrition
st.markdown("🔗 **Learn more about employee attrition from [Academy to Innovate HR (AIHR)](https://www.aihr.com/wp-content/uploads/Reasons-for-Employee-Attrition.png)**")

st.markdown("---")

# Main content
st.image("https://www.aihr.com/wp-content/uploads/Reasons-for-Employee-Attrition.png")

# Additional Information for Sample Prediction
st.write("📊 To make a sample prediction, you can refer to the training dataset information available in the sidebar or input the information of the employee whose attrition you want to predict.")
st.write("Please provide the following information to make a prediction:")

# About Section with Style
st.sidebar.title("ℹ️ About")
st.sidebar.info(
    "This app predicts employee attrition using machine learning on HR data, aiding HR professionals in retention strategies. "
    "It utilizes a CatBoost machine learning model trained on an employee attrition dataset."
)



# Define the Streamlit app
def main():
    # Define layout with three columns
    col1, col2, col3 = st.columns(3)

    # Column 1
    with col1:
        age = st.number_input("Age", min_value=18, max_value=70)
        monthly_income = st.number_input("Monthly Income")
        num_companies_worked = st.number_input("Number of Companies Worked")
        percent_salary_hike = st.number_input("Percent Salary Hike", min_value=0, max_value=25)
        training_times_last_year = st.number_input("Training Times Last Year", min_value=0, max_value=6)

    # Column 2
    with col2:
        department = st.selectbox("Department", ['Sales', 'Research & Development', 'Human Resources'])
        environment_satisfaction = st.selectbox("Environment Satisfaction", [1, 2, 3, 4])
        job_role = st.selectbox("Job Role", ['Sales Executive', 'Research Scientist', 'Laboratory Technician',
                                              'Manufacturing Director', 'Healthcare Representative', 'Manager',
                                              'Sales Representative', 'Research Director', 'Human Resources'])
        job_satisfaction = st.selectbox("Job Satisfaction", [1, 2, 3, 4])
        work_life_balance = st.selectbox("Work Life Balance", [1, 2, 3, 4])

    # Column 3
    with col3:
        over_time = st.checkbox("Over Time")
        relationship_satisfaction = st.selectbox("Relationship Satisfaction", [1, 2, 3, 4])
        years_since_last_promotion = st.number_input("Years Since Last Promotion")
        years_with_curr_manager = st.number_input("Years With Current Manager")

    # Predict button
    if st.button("Predict Employee Attrition 📊"):

        # Create a DataFrame to hold the user input data
        input_data = pd.DataFrame({
            'Age': [age],
            'Department': [department],
            'EnvironmentSatisfaction': [environment_satisfaction],
            'JobRole': [job_role],
            'JobSatisfaction': [job_satisfaction],
            'MonthlyIncome': [monthly_income],
            'NumCompaniesWorked': [num_companies_worked],
            'OverTime': [over_time],
            'PercentSalaryHike': [percent_salary_hike],
            'RelationshipSatisfaction': [relationship_satisfaction],
            'TrainingTimesLastYear': [training_times_last_year],
            'WorkLifeBalance': [work_life_balance],
            'YearsSinceLastPromotion': [years_since_last_promotion],
            'YearsWithCurrManager': [years_with_curr_manager]
        })

        # Reorder columns to match the expected order
        input_data = input_data[['Age', 'Department', 'EnvironmentSatisfaction', 'JobRole', 'JobSatisfaction',
                                 'MonthlyIncome', 'NumCompaniesWorked', 'OverTime', 'PercentSalaryHike',
                                 'RelationshipSatisfaction', 'TrainingTimesLastYear', 'WorkLifeBalance',
                                 'YearsSinceLastPromotion', 'YearsWithCurrManager']]

        # Make predictions
        prediction = model.predict(input_data)
        probability = model.predict_proba(input_data)[:, 1]
        
        # Display prediction probability
        if prediction[0] == 1:
            st.subheader("Prediction Probability 📈")
            st.write(f"The probability of the employee leaving is: {probability[0]*100:.2f}%")
        
        # Display characteristic-based recommendations
        st.subheader("Recommendations for Retaining The Employee 💡:")
        if job_satisfaction == 1 or environment_satisfaction == 1:
            st.markdown("- **Job and Environment Satisfaction**: Enhance job and environment satisfaction through initiatives such as recognition programs and improving workplace conditions.")
        if years_since_last_promotion > 5:
            st.markdown("- **Opportunities for Advancement**: Implement a transparent promotion policy and provide opportunities for career advancement.")
        if years_with_curr_manager > 5:
            st.markdown("- **Change in Reporting Structure**: Offer opportunities for a change in reporting structure to prevent stagnation and promote growth.")
        if percent_salary_hike < 5:
            st.markdown("- **Salary and Benefits Adjustment**: Consider adjusting salary and benefits packages to remain competitive and reward employee loyalty.")
        if training_times_last_year < 2:
            st.markdown("- **Employee Development**: Invest in employee development through training programs and continuous learning opportunities.")
        if over_time:
            st.markdown("- **Workload Evaluation**: Evaluate workload distribution and consider implementing measures to prevent overwork, such as workload balancing and flexible scheduling.")
        if relationship_satisfaction == 1:
            st.markdown("- **Positive Work Environment**: Foster positive relationships and a supportive work environment through team-building activities and open communication channels.")
        if monthly_income < 5000:
            st.markdown("- **Compensation Review**: Review compensation structures and adjust salaries to align with industry standards and employee expectations.")
        if num_companies_worked > 5:
            st.markdown("- **Address High Turnover**: Identify reasons for high turnover and address issues related to job stability, career progression, and organizational culture.")
        if work_life_balance == 1:
            st.markdown("- **Work-Life Balance Initiatives**: Promote work-life balance initiatives, such as flexible work arrangements and wellness programs, to support employee well-being.")

        # General recommendation for all negative predictions
        st.markdown("- **Exit Interviews**: Conduct exit interviews to gather feedback and identify areas for improvement in retention strategies.")

if __name__ == "__main__":
    main()