Question: Using the given data set, develop a cluster analysis. Remember to keep out the left variable from the cluster analysis, and to convert salary to a numerical variable.

Exploring Data

##  satisfaction_level last_evaluation  number_project  average_montly_hours
##  Min.   :0.0900     Min.   :0.3600   Min.   :2.000   Min.   : 96.0       
##  1st Qu.:0.4400     1st Qu.:0.5600   1st Qu.:3.000   1st Qu.:156.0       
##  Median :0.6400     Median :0.7200   Median :4.000   Median :200.0       
##  Mean   :0.6128     Mean   :0.7161   Mean   :3.803   Mean   :201.1       
##  3rd Qu.:0.8200     3rd Qu.:0.8700   3rd Qu.:5.000   3rd Qu.:245.0       
##  Max.   :1.0000     Max.   :1.0000   Max.   :7.000   Max.   :310.0       
##  time_spend_company Work_accident         left        promotion_last_5years
##  Min.   : 2.000     Min.   :0.0000   Min.   :0.0000   Min.   :0.00000      
##  1st Qu.: 3.000     1st Qu.:0.0000   1st Qu.:0.0000   1st Qu.:0.00000      
##  Median : 3.000     Median :0.0000   Median :0.0000   Median :0.00000      
##  Mean   : 3.498     Mean   :0.1446   Mean   :0.2381   Mean   :0.02127      
##  3rd Qu.: 4.000     3rd Qu.:0.0000   3rd Qu.:0.0000   3rd Qu.:0.00000      
##  Max.   :10.000     Max.   :1.0000   Max.   :1.0000   Max.   :1.00000      
##     sales              salary         
##  Length:14999       Length:14999      
##  Class :character   Class :character  
##  Mode  :character   Mode  :character  
##                                       
##                                       
## 

• Satisfaction level: Most employees are highly satisfied.
• Last Evaluation: Most employees are good performers with 75% of the data set being evaluated at 56 percent to 87 percent.
• Number of Projects: Most employees do a reasonable number of projects.
• Average Monthly Hours: Most employees spend fairly a higher amount of hours at work.
• Time Spent in Company: Fewer employees stay beyond 4 years.
  
• Work Accident: Most employees do not meet accidents at work.
• Left: Most employees stay with the organization and do not leave.
• Promotion in Last 5years: Most employees have not recieved promotions in the last five years.

Analyzing continous variables:

Analyzing categorical variables:

• Salary : 8.25 percent of the organization are top level with the highest pay, 42.9 percent of the employees are paid a medium salary and 48.7% of the employees are paid low salary.

Profile

Characterizing loyal employees:

• Employees who are paid low salaries leave the most, however, it is interesting to note that so do the employees with medium salaries. Ergo, a higher salary does not guarantee a loyal employee.
• The difference between the clusters who leave regardless of their pay grade is the promotion received in the past five years. Clusters with higher rate of leaving have also received fewer promotions and vice-versa.
• Employees who are least satisfied, as well as, who express a moderate level of satisfaction show a tendency to leave the organization. Ergo, satisfaction of the employees may not make for a loyal employee.
• The employees who left are good performers doing a high number of projects as a result of which they spend greater number of hours at work. Note that they received less promotions in the last five years on an average.
• The employees who left, stayed with the company for about 4-6years during which they met with fewer accidents.

Pre process data and selecting similarity measure.

Creating New Data Set for Cluster Analysis

## tibble [14,999 × 8] (S3: tbl_df/tbl/data.frame)
##  $ satisfaction_level   : num [1:14999] 0.38 0.8 0.11 0.72 0.37 0.41 0.1 0.92 0.89 0.42 ...
##  $ last_evaluation      : num [1:14999] 0.53 0.86 0.88 0.87 0.52 0.5 0.77 0.85 1 0.53 ...
##  $ number_project       : num [1:14999] 2 5 7 5 2 2 6 5 5 2 ...
##  $ average_montly_hours : num [1:14999] 157 262 272 223 159 153 247 259 224 142 ...
##  $ time_spend_company   : num [1:14999] 3 6 4 5 3 3 4 5 5 3 ...
##  $ Work_accident        : num [1:14999] 0 0 0 0 0 0 0 0 0 0 ...
##  $ promotion_last_5years: num [1:14999] 0 0 0 0 0 0 0 0 0 0 ...
##  $ salary_num           : num [1:14999] 10 20 20 10 10 10 10 10 10 10 ...

• To comply with the requirements of the question, a new data set called ‘salary’ was created by mutating alphanumeric variable ‘salary’ to the one with numerical values.
• The variable ‘sales’ was de-slected and the ‘salary’ data set produced contained 9 variables for 14999 observations.
• Data set ‘salary_ca’ was created where the variable ‘left’ was deselected. It is this data set which will be standardized and scaled for accruately evaluating the profiles of the employees.

Creating Clusters using K-means clustering

• Standardization and scaling were employed with the purpose of normalizing the raw data values and assigning a singular format to numbers for accurate analysis.
• Optimal number of K means clusters for the data set were decided by the method of total within-cluster sum of squares (wss).
• 8 clusters were decided based on the logic that the measure of compactness on the plot produced post standardization and scaling is minimal beyond 8.

Profiling and Plotting:

Analysis

Cluster 8 has the maximum observations which left.

• Cluster 8 has one of the lowest number of observations amounting to 872, of which, 65 percent left.
• Cluster 8 is the least satisfied. _ This cluster is paid the lowest salary.
• The employees of this cluster are very good performers with an evaluation of 82 percent on an average.
• This cluster has employees who do the highest number of projects, so, end up spending the highest amount of time at work on an average on a monthly basis.
• Employees in this cluster tend to stay for about 4 years on an average with the . Ergo, they are not new employees.
• They have received extremely few promotions in the last five years.
• It is fair to conclude that cluster 8, on an average, is representative of workers in this organization.

Cluster 5 has the highest number of observations.

• Cluster 5 has 3,834 observations which are highly salaried.
• They express a higher level of satisfaction.
• They are very good performers evaluated at approximately 73 percent.
• They do a moderate number of projects spending a relatively higher number of hours working on an average on a monthly basis.
• Of the employees represented in this cluster, 1.7 percent left the organization.
• The employees in this cluster spend approximately 3years in the organization during on an average have met no accidents.
• Approximately 31 percent of the observations in this cluster have received promotion in the last 5 years.
• It is fair to conclude that on an average, this cluster, represents managers and executives.

Cluster 5 is paid the highest salary.

• On an average, cluster 5 is paid the highest salary and expresses a higher satisfaction level. Since, it comprises of approximately 25 percent of the employees, it is fair to conclude that this cluster largely represents managers.
  
• Clusters 2, 3, 4 and 6 are paid medium salaries for a moderate performance on working a moderate number of projects who stay with the company for 3-6years, do not meet accidents and have recieved moderate amount of promotions in the last five years.
• Clusters 7 is paid the lowest salary, however, they are highly satisfied and a small percent of which has left. They are good performers who work on a moderate number of projects and tend to leave the organization by approximately 3 years. They receive extremely low level of promotions. This cluster is representative of workers.

Cluster 8 does a maximum number of projects.

• Cluster 8 does the highest number of projects. Besides which, cluster 1 does the most projects.
• Cluster 1 is the smallest cluster which is paid a high salary.
• Also, cluster 1 who are very good performers express low satisfaction for working on a higher number of projects.
• Cluster 1 represents observations which do not tend to stay longer than 4years and during which, 1.1 percent have received promotions in the last five years.

Cluster 8 spends the maximum number of hours on an average on a monthly basis.

• Since clusters 8 and 1 do the maximum number of projects, they tend to be spending the most time in the office on an average on monthly basis.
• Similarly, cluster 4 which tends be doing least number of projects end up spending the least amount of hours working.

Clusters 2 and 6 suffer the most accidents at work.

• Cluster 2 and 6 meet the most accidents at work.
• It is important to note that besides this, these both clusters are paid medium salaries and relatively, spend a fair amount of time doing moderate number of projects.
• In cluster 2 approximately 2.1 percent and in 6, 4.7 percent (highest) have received promotion in the last five years.
• Of the total 939 observations, 84 percent in cluster 2 have left.
• Of the total 1089 observations, 46 percent in cluster 6 have left.

Cluster 3 has the highest level of satisfaction.

• Cluster 3 has 1,624 observations which express the highest satisfaction level at 75 percent.
• They spend a moderate amount of hours working on a moderate number of projects on a medium level salary.
• They are evaluated as good performers who tend to stay with the organization for 6 years on an average.
• Of the total, 3.1 percent have received promotions in the last five years.
• Approximately 51 percent observations of this cluster have left the organization.

Cluster 3 spends most time with the company.

• Cluster 3 has stayed the longest time with the company even though 51 percent of its employees have left.
• This aligns well with the observation that they are also the most satisfied.

Cluster 6 has experienced highest number of promotions in the last five years.

• Cluster 6 has 1,089 observations of which 4.7 percent have received promotions in the last five years working on a medium level salary.
• They express a moderate level of satisfaction even though they meet the most accidents at work and spend a relatively larger number of hours working on an average on monthly basis.
• Only 4.6 percent of the employees in cluster 6 have left the organization.

Cluster 3 has received the highest number of last evaluations.

• Cluster 3 has 1,624 observations and are evaluated at 83 percent which tells that they are very good performers.
• CLuster 3 expresses the highest level of satisfaction and only 3.6 percent of the employees in the cluster have left.
• They also stay the longest with the company.