EDA of HR Dataset

Dataset Used

Raw dataset used during analysis:
Link to dataset: HRDataset_v14.xlsx

---

Initial Setup

• Imported required Python libraries for EDA.
• Loaded the dataset and converted it into a DataFrame for analysis.
• Checked basic dataset information:
  • Number of rows and columns
  • Data types of all columns
  • Missing values (only the column "DateofTermination" had 207 null values)
    • Replaced missing values with '0' instead of dropping the column
  • Checked for duplicate values (none found)

---

Analysis Performed

• Top 10 employees receiving the highest salaries
• Identified underperforming employees who need assistance
• Listed employees with high leave counts
• Analyzed marital status of employees
• Checked which employees are involved in special projects
• Created bar charts to compare top 10 and lowest 10 salaries
• Analyzed source of recruitment using horizontal bar charts
• Analyzed performance scores by categories: Full Meets, Exceeds, Needs Improvement, and PIP
• Performed multivariate analysis:
  • Boxplot to detect salary outliers by department
  • Relationship between Engagement Survey Score and Position
  • Marital status breakdown by gender
  • Average engagement survey score by department
• Termination analysis:
  • Total number of terminated employees
  • Terminations by department and by position
• Median salary by gender
• Analyzed reasons for employee termination
• Maximum number of absences by department
• Total absences and average engagement score for each department
• Number of terminated employees and their average satisfaction score by department
• Special projects count and average absences by gender

---

Conclusion

This analysis provides a comprehensive overview of employee performance, satisfaction, and attrition trends in the HR dataset, which can help with better decision-making.