1a. Create a boxplot for employee satisfaction, split by ‘left’ .

#Employee satisfaction is crucial for retention efforts, as the boxplot is likely to show lower satisfaction levels among departing employees than among remaining ones.

1b. Create box plots for last evaluation, split by ‘left’.

#Employees who left (left = 1) may have received lower final evaluation scores than those who stayed (left = 0), as the boxplot is likely to demonstrate. This may indicate that recent evaluations have a big impact on keeping employees on board, meaning performance reviews should be given careful consideration when developing employee satisfaction and retention plans.

2.Select Continuous Variables

##                      satisfaction_level last_evaluation number_project
## satisfaction_level           1.00000000       0.1050212     -0.1429696
## last_evaluation              0.10502121       1.0000000      0.3493326
## number_project              -0.14296959       0.3493326      1.0000000
## average_montly_hours        -0.02004811       0.3397418      0.4172106
## time_spend_company          -0.10086607       0.1315907      0.1967859
##                      average_montly_hours time_spend_company
## satisfaction_level            -0.02004811         -0.1008661
## last_evaluation                0.33974180          0.1315907
## number_project                 0.41721063          0.1967859
## average_montly_hours           1.00000000          0.1277549
## time_spend_company             0.12775491          1.0000000

Create the correlogram using corrplot

#An extensive summary of the relationships between the continuous variables in your dataset can be obtained from the correlogram. It enables the rapid identification of strong relationships—such as those influencing employee satisfaction or performance reviews—that may call for additional research or modeling consideration. Furthermore, the visualization aids in emphasizing possible problems with multicollinearity that might occur when incorporating these variables into predictive models.