Client is a U.S. Healthcare provider which specializes in hospital services, mental health services, women’s health services, specialty services, and substance abuse work in the community to deliver a healthcare experience designed around individual needs.

They were looking to revolutionize their performance evaluation system and build a solution that could analyze employee data, predict employee performance and offer unbiased decision-making capabilities & recommendations to improve employee performance.

ClientHealthcareServicesData Science, AI, MLYear 2022

Key Challenges

  • The healthcare client wanted to overcome issues found in their current evaluation system as it was dealt with by humans​​​​
  • Humans default to their emotions, biases, prejudices, etc., which could negatively impact the organization’s growth and lead to inefficient decisions ​
  • Additionally, managing and analyzing vast quantities of HR data manually is time-consuming and prone to errors
  • The client needed a solution to help them eliminate inefficiencies that lower employee morale with a relatively easy-to-employ framework


  • Sparity developed and implemented predictive models for performance evaluation using machine learning algorithms
  • Employed Hybrid techniques and developed and trained two predictive models, Data clustering and Decision Tree Classifier algorithms ​
  • Applied data clustering for evaluating the employee’s performance and decision-making process and imported python libraries — NumPy and pandas
  • The predictive model considered different performance evaluation factors like personality, punctuality, tact, oral expression, Quality of Work, perseverance, public relations, observance of security measures, capacity to guide & train subordinates, attitude towards superiors, moral integrity, and more to provide actionable insights and predict and evaluate the performance of the employees ​ ​
  • Used the Label encoder utility class from scikit-learn and transformed those diverse categorical values into numerical ​
  • Leveraged Decision tree classifier to help visualize and analyze the situation better and evaluate the employee’s performance as well as in the decision-making process ​ ​


  • Attained maximum productivity by reducing manual workload by more than 35%
  • Empowered client to identify over and underutilized resources and design relevant training courses for a specific period of time​
  • Minimized the bias that’s inherited in the old-school method of performance management
  • Predicts the employees’ performance for the following year and enabled easy advancement and promotion determinations