Case Study - Employee Attrition for Healthcare Using R and PowerBI

Introduction

This exploratory analysis case study is towards Capstone project requirement for Google Data Analytics Professional Certificate. The case study involves a Dataset of Employee Attrition for Healthcare.

The analysis will follow the 6 phases of the Data Analysis process: Ask, Prepare, Process, Analyze, and Act (APPAA).

ASK

Task: Analyze employee attrition rates by department, educational field, age group, and gender to identify patterns and to help HR develop retention strategies.

Description: To address the issue of employee attrition, we need to analyze the attrition rates of each department, educational field, age group, and gender to identify patterns and develop targeted retention strategies.

These are the questions that needs to be answered.

• What is the attrition rate of each department?
• What is the attrition rate based on educational field?
• What is the number of employees by age group?
• What is the attrition rate by gender for different age groups?

PREPARE

Data Source

ABOUT THE DATASET:

This dataset contains employee and company data useful for the analysis. Attrition - whether an employee left or not - is included and can be used as the target variable. The data is synthetic and based on the IBM Watson dataset for attrition. Employee roles and departments were changed to reflect the healthcare domain.

The Dataset is from Employee Attrition for Healthcare. The data has been made available by John Miller under CC0: Public Domain license.

PROCESS

Documentation, Cleaning and Preparation of data for analysis

Install and load R libraries

install.packages("tidyverse")
install.packages("skimr")
install.packages("dplyr")
library(tidyverse)
library(skimr)
library(dplyr)

     

Load Dataset

attrition_data <- read_csv("...\\watson_healthcare_modified.csv")

     

Inspect the Data Frame

head(attrition_data)
glimpse(attrition_data)
View(attrition_data)

PROCESS

Data Cleaning before conducting analysis

Delete duplicate rows

attrition_data_unique <- distinct(attrition_data)

     

Check for missing values

colSums(is.na(attrition_data))

Adding new columns that can be used for aggregate functions

Create attrition count column

attrition_data$Attrition_Count <- as.integer(ifelse(attrition_data$Attrition == "Yes", 1, 0))

     

Calculate the sum of Attrition_Count by department

dept_counts <- aggregate(Attrition_Count ~ Department, data = attrition_data, FUN = sum)

     

Calculate the percentage of Attrition_Count by department

dept_counts$Percent <- dept_counts$Attrition_Count / sum(dept_counts$Attrition_Count) * 100

     

Create a new column 'Age_Band'

attrition_data$Age_Band <- ifelse(attrition_data$Age >= 55, 'Over 55',
                                  ifelse(attrition_data$Age >= 45, '45-54',
                                         ifelse(attrition_data$Age >= 35, '35-44',
                                                ifelse(attrition_data$Age >= 25, '25-34',
                                                       'Under 25'))))

ANALYZE

Let's check the attrition by Age Band

summary_data1 <- attrition_data %>%
  group_by(Age_Band) %>%
  summarise(
    Attrition_Count = sum(Attrition_Count),
    EmployeeCount = n(),
  )
print(summary_data1)

     

Let's check the attrition by Department

summary_data2 <- attrition_data %>%
  group_by(Department) %>%
  summarise(
    Attrition_Count = sum(Attrition_Count),
    EmployeeCount = n(),
  )
print(summary_data2)

     

Let's check the attrition by Education Field

summary_data3 <- attrition_data %>%
  group_by(EducationField) %>%
  summarise(
    Attrition_Count = sum(Attrition_Count),
    EmployeeCount = n(),
  )
print(summary_data3)

     

Visualizations

Attrition by Age Band

age_band_sorted <- attrition_data %>%
  mutate(Age_Band = factor(Age_Band, 
                                 levels = c("Under 25", "25-34", "35-44", "45-54", "Over 55")))

ggplot(data = age_band_sorted, aes(x = Age_Band, y = Attrition_Count)) + 
  geom_bar(stat = "identity") +
  labs(x = "Age band", y = "Attrition count", title = "Attrition by Age Band") +
  theme(plot.title = element_text(hjust = 0.5))

     

• Findings: The age band with the highest attrition count is 25-34, with 106 employees leaving the organization, which represents 16.67% of the total employees in that age group (106/636)

     

Attrition by Department

ggplot(dept_counts, aes(x = "", y = Attrition_Count, fill = Department)) +
  geom_bar(width = 1, stat = "identity") +
  coord_polar(theta = "y") +
  geom_text(aes(label = paste0(Attrition_Count, "\n", paste0("(", round(Percent, 2), "%", ")"))),
            position = position_stack(vjust = 0.5)) +
  labs(fill = "Department", title = "Attrition Counts by Department (Pie Chart)") +
  theme_void() +
  theme(legend.position = "right",
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        panel.grid = element_blank(),
        plot.title = element_text(hjust = 0.5, size = 18),
        legend.title = element_text(size = 12),
        legend.text = element_text(size = 10))

     

• Findings: The highest attrition count is in the Maternity department with 98 employees leaving the organization, followed by Cardiology with 74 employees leaving, and then Neurology with 27 employees leaving.

     

Attrition by Education Field

EducationField_sorted <- attrition_data %>%
  mutate(EducationField = factor(EducationField, 
                                 levels = c("Human Resources", "Other", "Technical Degree", "Marketing", "Medical", "Life Sciences")))

ggplot(EducationField_sorted, aes(x = Attrition_Count, y = EducationField)) +
  geom_bar(stat = "identity") +
  labs(title = "Attrition Count by Education Field", x = "Attrition Count", y = "") +
  theme_minimal() +
  theme(plot.title = element_text(hjust = 0.5))

• Findings: The Education Field with the highest attrition count is Life Sciences with 84 employees leaving the organization, followed by Medical with 51 employees leaving, and then Marketing with 28 employees leaving. Among the Education Fields, Human Resources has the lowest attrition count with only 6 employees leaving, followed by Other with 8 employees leaving, and then Technical Degree with 22 employees leaving.

     

Additional Visualization

DYNAMIC DASHBOARD THRU POWER BI

I leveraged Power BI to design and develop a dynamic dashboard that presents key data in an interactive and visually appealing format. To see it in action click here!

ACT

Conclusion:

The organization has a significant issue with employee attrition, particularly in the Maternity, Cardiology, Life Sciences, and Medical Education departments. Additionally, younger employees (25-34) are leaving the organization at a higher rate than other age groups.

Recommendations:

1. Maternity and Cardiology departments have the highest attrition counts. HR should conduct an analysis to identify the root causes of the high attrition rates, and take appropriate actions to address those issues. For example, HR can implement retention programs, offer flexible work arrangements, provide adequate training and development opportunities, and improve communication with employees.

2. Life Sciences and Medical Education Fields have the highest attrition counts. HR should analyze the reasons behind the high attrition rates in these fields, and take measures to address the underlying causes. For instance, HR can provide more opportunities for career development, offer competitive compensation packages, and foster a positive work environment.

3. Human Resources, Other, and Technical Degree Education Fields have the lowest attrition counts. HR should investigate the reasons behind the low attrition rates in these fields and leverage them to improve employee retention in other departments and education fields. HR can identify and implement best practices, such as providing a supportive work environment, offering opportunities for learning and growth, and recognizing employee achievements.

4. Younger employees (25-34) have the highest attrition rate. HR should focus on creating an inclusive and supportive work environment, providing career advancement opportunities, and offering competitive compensation packages to retain younger employees. Additionally, HR should conduct an analysis to identify the factors that drive younger employees to leave the organization, and implement appropriate measures to address those issues.

By taking these actions, the organization can improve employee retention, reduce attrition rates, and foster a positive and productive work environment for all employees.

References

John Miller for the dataset. Thank you for providing a great dataset!  

Dataset  

Dataset License: CC0: Public Domain

Contact Information

LinkedIn  

UpWork