EmployeeAttrition

• Obtain employee data maintained by the HR department of a company. This data was extracted from Kaggle.
• Perform data cleaning, manipulation and pre-processing using Python and SQL.
• Perform univariate, bivariate and multivariate analysis to identify various factors leading to employee attrition.
• Use Tableau to create interactive visualizations and dashboards to track attrition metrics, and analyze latest trends.
• Suggest employee retention methods by targetting issues that contribute heavily to employee attrition.

Data Analysis

The process of analysis was broken down into following sections:

• Univariate analysis → To summarize the main characteristics of each attribute and describe its distribution, central tendency, dispersion, and shape.
• Bivariate analysis → To determine if there is a relationship or association between each independent variable, and the dependent variable ‘turnover’
• Multivariate analysis → To understand the relationships and dependencies among important features and the dependent variable ‘turnover’

Univariate analysis

• There are 12000 employees in 'XYZ Corp', and 2000 have left the company in the last year (almost 17% attrition rate). The company has 10 departments, with 'sales', 'technical', and 'support' departments comprising 61% of the total employees.
• The average satisfaction level is around 62%, with 58% of the employees highly satisfied, and 12% having low satisfaction. This tells us that people are generally satisfied working in this company, but there is still room for improvement as 42% of the employees feel some level of dissatisfaction.

• The average years that people stay in the company is 3.36 years, with the majority of employees spending 2-3 years. However there is a high variation in this value, with minimum value of 2 years and maximum of 10 years.
• The company has high performers, with 53% of the employees getting a high evaluation (> 0.7) last year. The performance average for the company is around 71%. This is an area that needs attention - are high performers leaving the company? Are they satisfied with the company?
• The salary range is highly skewed, with only 8% of the employees falling under the high salary range. This could be due to the fact that generally high compensation is offered to select few positions - the upper management and senior technical managers. However, with 53% of employees receiving very high evaluations, this number seems to be small. Could this be a reason for dissatisfaction among employees? Should the salary offering of the company be revised to offer better compensation to high performers.

• 'Average monthly hours' for this company is quite high, where 71% of the employees work for more than the standard working hours (160 hours monthly). It follows a bi-modal distribution, peaking at around 160 hours and 260 hours.

• Promotion is another factor that seems worrying as less than 3% of the employees received promotion in the last 5 years. The promotion policy demands a revision.

Bivariate analysis

• As expected, there is a greater turnover percentage among employees with lesser satisfaction (less than 0.5).

• However, we do have 550 employees who were highly satisfied (more than 0.7) who left the company. This is alarming. Of all the employees who left, more than 25% were highly satisfied. These employees' exit interview feedback should be observed closely to know the reason behind their attrition.

• Maximum attrition is seen among people who have been in the company for 3-5 years. This could be due to factors like:
  • Low salary growth
  • Not finding the job challenging

• The mean evaluation of turnover employees is higher than that of current employees. This means that employees who left were on average better performers than those who stayed.

• Barring just 1 employee, every other employee who left did not receive promotion in the last 5 years. This is compelling evidence that lack of promotion, among other factors, is a big reason that employees leave.

Multivariate Analysis

From the above analysis, there are two important issues that stand out and need further analysis:

1. The mean evaluation rating of turnover employees is higher than that of retained employees. Why did good performers leave?

2. 25% of the turnover employees were highly satisfied. Why did they leave?

Performance wise turnover

The performance wise turnover paints an interesting picture. Only those employees left the company who received a ‘Meets Expectation’ or a ‘Exceeds Expectation’ rating in the last evaluation.

Also, among employees who received a very high rating (Exceeds Expectation), the majority of them were working for long hours.

This is a serious issue that needs immediate redressal. Good performers are leaving the company due to overwork, whereas low performers continue to work in the company. In fact, out of 2000 turnovers, 1042 were top performers, and overworked:

Satisfaction wise turnover

To understand why highly satisfied people leave the company, it was important to identify the top factors affecting attrition. Following methods were used to achieve this:

• Visual plots
• ANOVA test: To identify important ‘numerical’ attributes affecting attrition

Since ANOVA is used for selecting NUMERICAL features, we'll ignore the CATEGORICAL features from the above list. Following NUMERICAL features have highest correlations with the dependent variable 'turnover'. 1. satisfaction_level 2. time_spend_company 3. average_montly_hours

• Chi-square test: To identify important ‘categorical’ attributes affecting attrition

Since Chi-square is used for selecting CATEGORICAL features, we'll ignore the NUMERICAL features from the above list. Following CATEGORICAL features have good correlation with the dependent variable 'turnover'. 1. Work_accident 2. Salary

• features_importances_ : This method gives weighted importance of all the features that the Machine Learning model used to make predictions for the given dataset.

Using the features_importances_ report of the Random Forest model, among all the factors, satisfaction_level’, ‘time_spend_company’ and ‘average_monthly_hours’ were identified to be the biggest factors leading to employee turnover.

Analyzing these 3 factors together:

The above plot clearly shows that highly satisfied employees who left were working long hours, and had been in the company for 5-6 years.

Also, there seems to be a level of dissatisfaction among employees who were in the company for 3 to 4 years.

Analyzing salary range for this group:

Employees who were in the company for 3-4 years had a low-medium salary range.

Key takeaways and Potential solutions:

• Looking at the data, following are the key points:
• It seems that many people are leaving because of low levels of satisfaction, not getting promoted and over-work.
• Insufficient compensation is a big reason for dissatisfaction among employees.
• A common factor among most turnover employees is lack of promotion, due to which employees do not find their jobs rewarding, leading to their attrition.
• Highest attrition is in the HR department, followed by Accounting and Technical teams.

Based on the above identified factors, following solutions should be put in place:

• Restrict work hours → The HR policy needs a change to implement stricter working hours, not exceeding the expected average of 160 hours per month.
• Better compensation → A common factor among dissatisfied employees was low to medium salary ranges. The company can follow industry standards as criteria to drive its cost to company and salary guidelines.
• Promotion policy review → With more than 99% of turnover employees not getting promoted in 5 years, the promotion policy needs an immediate review to address the extremely low rate of promotion in the company. Department oriented → Efforts should be more focussed on departments seeing higher attrition rates - HR, Accounting and Technical.