Employee Growth Score Predictor 📈

A Gradient Boosting Regression model that predicts an employee's growth score (0-100) based on their HR profile features.

Model Description

This model predicts a composite "Employee Growth Score" that captures an employee's overall career growth trajectory. The score is engineered from multiple HR dimensions:

Component	Weight	Description
Promotion Velocity	25%	How frequently the employee gets promoted
Job Level Achievement	20%	Current job level (1-5 scale)
Income-to-Experience Ratio	20%	Earning power relative to years of experience
Satisfaction Composite	15%	Combined job + environment satisfaction
Job Involvement	10%	Level of engagement and involvement
Retention Signal	10%	Whether the employee is retained (vs. attrited)

Performance Metrics

Metric	Value
Test R²	0.8997
Test RMSE	4.0360
Test MAE	2.9716
CV R² (mean ± std)	0.8631 ± 0.0194

Feature Importances

Feature	Importance
YearsSinceLastPromotion	0.3244
MonthlyIncome	0.2196
YearsAtCompany	0.0937
JobLevel	0.0893
JobSatisfaction	0.0704
EnvironmentSatisfaction	0.0633
JobInvolvement	0.0574
TotalWorkingYears	0.0316
DailyRate	0.0122
Age	0.0098
YearsInCurrentRole	0.0076
DistanceFromHome	0.0066
MaritalStatus_Num	0.0048
JobRole_Num	0.0042
YearsWithCurrManager	0.0022
BusinessTravel_Num	0.0021
Department_Num	0.0008

Usage

import joblib
import numpy as np
from huggingface_hub import hf_hub_download

# Download and load model
model_path = hf_hub_download(
    repo_id="dev91205/employee-growth-score-predictor",
    filename="employee_growth_model.joblib"
)
model = joblib.load(model_path)

# Feature order: Age, BusinessTravel_Num, DailyRate, Department_Num,
# DistanceFromHome, EnvironmentSatisfaction, JobInvolvement, JobLevel,
# JobRole_Num, JobSatisfaction, MaritalStatus_Num, MonthlyIncome,
# TotalWorkingYears, YearsAtCompany, YearsInCurrentRole,
# YearsSinceLastPromotion, YearsWithCurrManager

employee = np.array([[30, 2, 800, 2, 5, 3, 3, 2, 5, 3, 1, 5000, 8, 5, 3, 1, 3]])
score = model.predict(employee)
print(f"Growth Score: {score[0]:.1f}/100")

Growth Score Categories

Score Range	Category	Description
75-100	🟢 HIGH GROWTH	Strong career trajectory, ready for advancement
50-74	🟡 MODERATE GROWTH	Solid performance, potential for development
25-49	🟠 LOW GROWTH	Needs attention, potential stagnation
0-24	🔴 AT RISK	Requires immediate development support

Training Details

Dataset: eduvance/employee_attrition
Algorithm: Gradient Boosting Regressor (scikit-learn)
Training samples: 959
Test samples: 411
Hyperparameter tuning: GridSearchCV with 5-fold cross-validation (384 candidates)

Best Hyperparameters

Parameter	Value
learning_rate	0.05
max_depth	3
min_samples_split	10
n_estimators	500
subsample	0.8

Limitations

Based on IBM HR Analytics-style data — may not generalize to all industries
Growth score is an engineered composite, not a ground-truth label
Does not account for external factors (market conditions, company growth stage)
Feature encoding is pre-computed (categorical → numeric)

Downloads last month: -

dev91205
/

employee-growth-score-predictor