Logistic regression for hr attrition case study

1. Submitted by group 1
Laxmi S Kallimath
CASE STUDY ON HR ATTRITION
RATE

2. OBJECTIVE
• TO DETERMINE THE FACTORS CONTRIBUTING TO
EMPLOYEE ATTRITION AND THEIR EXTENT

3. DATA
35 VARIABLES
1470
OBSERVATIONS
 BusinessTravel
 DistanceFromHome
 EnvironmentSatisfaction
 JobInvolvement
 OverTime
 YearsInCurrentRole
 JobSatisfaction
 MaritalStatus
 WorkLifeBalance
 JobRole

4. Data Set
Age Attrition BusinessTravel DailyRate Department DistanceFromHome Education EducationField EmployeeCount EmployeeNumber EnvironmentSatisfaction Gender HourlyRate
41 Yes Travel_Rarely 1102 Sales 1 2 Life Sciences 1 1 2 Female 94
49 No Travel_Frequently 279 Research & Development 8 1 Life Sciences 1 2 3 Male 61
37 Yes Travel_Rarely 1373 Research & Development 2 2 Other 1 4 4 Male 92
33 No Travel_Frequently 1392 Research & Development 3 4 Life Sciences 1 5 4 Female 56
27 No Travel_Rarely 591 Research & Development 2 1 Medical 1 7 1 Male 40
32 No Travel_Frequently 1005 Research & Development 2 2 Life Sciences 1 8 4 Male 79
59 No Travel_Rarely 1324 Research & Development 3 3 Medical 1 10 3 Female 81
30 No Travel_Rarely 1358 Research & Development 24 1 Life Sciences 1 11 4 Male 67
38 No Travel_Frequently 216 Research & Development 23 3 Life Sciences 1 12 4 Male 44
36 No Travel_Rarely 1299 Research & Development 27 3 Medical 1 13 3 Male 94
35 No Travel_Rarely 809 Research & Development 16 3 Medical 1 14 1 Male 84
29 No Travel_Rarely 153 Research & Development 15 2 Life Sciences 1 15 4 Female 49
31 No Travel_Rarely 670 Research & Development 26 1 Life Sciences 1 16 1 Male 31
34 No Travel_Rarely 1346 Research & Development 19 2 Medical 1 18 2 Male 93
28 Yes Travel_Rarely 103 Research & Development 24 3 Life Sciences 1 19 3 Male 50
29 No Travel_Rarely 1389 Research & Development 21 4 Life Sciences 1 20 2 Female 51
32 No Travel_Rarely 334 Research & Development 5 2 Life Sciences 1 21 1 Male 80
22 No Non-Travel 1123 Research & Development 16 2 Medical 1 22 4 Male 96
53 No Travel_Rarely 1219 Sales 2 4 Life Sciences 1 23 1 Female 78
38 No Travel_Rarely 371 Research & Development 2 3 Life Sciences 1 24 4 Male 45
24 No Non-Travel 673 Research & Development 11 2 Other 1 26 1 Female 96
36 Yes Travel_Rarely 1218 Sales 9 4 Life Sciences 1 27 3 Male 82
34 No Travel_Rarely 419 Research & Development 7 4 Life Sciences 1 28 1 Female 53
21 No Travel_Rarely 391 Research & Development 15 2 Life Sciences 1 30 3 Male 96
34 Yes Travel_Rarely 699 Research & Development 6 1 Medical 1 31 2 Male 83
53 No Travel_Rarely 1282 Research & Development 5 3 Other 1 32 3 Female 58
32 Yes Travel_Frequently 1125 Research & Development 16 1 Life Sciences 1 33 2 Female 72
42 No Travel_Rarely 691 Sales 8 4 Marketing 1 35 3 Male 48
44 No Travel_Rarely 477 Research & Development 7 4 Medical 1 36 1 Female 42
46 No Travel_Rarely 705 Sales 2 4 Marketing 1 38 2 Female 83
33 No Travel_Rarely 924 Research & Development 2 3 Medical 1 39 3 Male 78
44 No Travel_Rarely 1459 Research & Development 10 4 Other 1 40 4 Male 41
30 No Travel_Rarely 125 Research & Development 9 2 Medical 1 41 4 Male 83
39 Yes Travel_Rarely 895 Sales 5 3 Technical Degree 1 42 4 Male 56
24 Yes Travel_Rarely 813 Research & Development 1 3 Medical 1 45 2 Male 61
43 No Travel_Rarely 1273 Research & Development 2 2 Medical 1 46 4 Female 72
50 Yes Travel_Rarely 869 Sales 3 2 Marketing 1 47 1 Male 86
35 No Travel_Rarely 890 Sales 2 3 Marketing 1 49 4 Female 97
36 No Travel_Rarely 852 Research & Development 5 4 Life Sciences 1 51 2 Female 82
33 No Travel_Frequently 1141 Sales 1 3 Life Sciences 1 52 3 Female 42
35 No Travel_Rarely 464 Research & Development 4 2 Other 1 53 3 Male 75
27 No Travel_Rarely 1240 Research & Development 2 4 Life Sciences 1 54 4 Female 33
26 Yes Travel_Rarely 1357 Research & Development 25 3 Life Sciences 1 55 1 Male 48
27 No Travel_Frequently 994 Sales 8 3 Life Sciences 1 56 4 Male 37
30 No Travel_Frequently 721 Research & Development 1 2 Medical 1 57 3 Female 58
41 Yes Travel_Rarely 1360 Research & Development 12 3 Technical Degree 1 58 2 Female 49
34 No Non-Travel 1065 Sales 23 4 Marketing 1 60 2 Male 72
37 No Travel_Rarely 408 Research & Development 19 2 Life Sciences 1 61 2 Male 73
46 No Travel_Frequently 1211 Sales 5 4 Marketing 1 62 1 Male 98
35 No Travel_Rarely 1229 Research & Development 8 1 Life Sciences 1 63 4 Male 36
48 Yes Travel_Rarely 626 Research & Development 1 2 Life Sciences 1 64 1 Male 98
28 Yes Travel_Rarely 1434 Research & Development 5 4 Technical Degree 1 65 3 Male 50
44 No Travel_Rarely 1488 Sales 1 5 Marketing 1 68 2 Female 75
35 No Non-Travel 1097 Research & Development 11 2 Medical 1 70 3 Male 79
26 No Travel_Rarely 1443 Sales 23 3 Marketing 1 72 3 Female 47
33 No Travel_Frequently 515 Research & Development 1 2 Life Sciences 1 73 1 Female 98
35 No Travel_Frequently 853 Sales 18 5 Life Sciences 1 74 2 Male 71
35 No Travel_Rarely 1142 Research & Development 23 4 Medical 1 75 3 Female 30
31 No Travel_Rarely 655 Research & Development 7 4 Life Sciences 1 76 4 Male 48
37 No Travel_Rarely 1115 Research & Development 1 4 Life Sciences 1 77 1 Male 51
32 No Travel_Rarely 427 Research & Development 1 3 Medical 1 78 1 Male 33
38 No Travel_Frequently 653 Research & Development 29 5 Life Sciences 1 79 4 Female 50

5. METHODOLOGY
IMPORT THE DATA INTO R
CHECK THE IMPORTED DATA USING
STRUCTUREDIMENSIONSUMMARYNAMES
HEAD FUNCTIONS
NO MISSING VALUES AND NO NA VALUES
IDENTIFY DEPENDENT VARIABLE – ATTRITION
ALL OTHER VARIABLES ARE INDEPENDENT

6. ATTRITION – YESNO
SINCE WE ARE USING LOGISTIC REGRESSION WE
CONVERT YES NO INTO BINARY FORM
ATTRITION YES = 1
ATTRITION NO = 0
CONVERT CATEGORICAL VARIABLES INTO
FACTORS

7. UNIVARIATE AND BIVARIATE
ANALYSIS
• MEAN
• MEDIAN
• STANDARD DEVIATION
• MIN
• MAX
• SUMMARY
UNIVARIATE
• DIVIDE DATA INTO NUMERIC AND
CATEGORICAL
• NUMERIC – INDEPENDENT T TEST
• CATEGORICAL – CHI SQUARE
TEST
BIVARIATE
(TO IDENTIFY
SIGNIFICANT AND
NON SIGNIFICANT
VARIABLES)

8. BIVARIATE ANALYSIS
• H0 : MEAN(Age(0)) = MEAN(Age(1))
• H1: MEAN(Age(0)) = MEAN(Age(1))

9. Interpretation:-
 p-value = 1.38e-08 p<0.05
Reject Ho Age is significant

10. Chi Square Test (Job Role)

11. Chi Square Test(marital Status)

12. DATA SPLIT
• HERE DATA IS BALANCED
• ATRRITION YES (1) – 16%
• ATRRITION NO (0) - 84%
• USING SIMPLE RANDOM SAMPLING WITHOUT
REPLACEMENT
• SPLIT THE DATA INTO TEST AND TRAIN DATA SET
• TRAIN – 70%
• TEST – 30%

13. Model Output
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

14. Check assumption
• Multi collinearity
• Vif < 5 no multi collinearity

15. validation
 Sensitivity = True Positive / (True Positive False Negative)
 Specificity = True Negative / (True Negative False Positive)
 Accuracy = (true positive true negative) / (total sum of all
observations)
 TRAIN DATA SET
Actual/predicted No(0) Yes(1)
0 823 29
1 106 71
Sensitivity = 40%
Specificity = 96%
Accuracy = 86%

16. • TEST DATA SET
Actual/predicted No(0) Yes(1)
0 367 14
1 34 26
Sensitivity = 43%
Specificity = 96%
Accuracy = 89%

17. TRAIN
CONDORDANCE = 84%
DICORDANCE = 15%
PAIRS = 150804
TEST
CONDORDANCE = 82%
DICORDANCE = 18%
PAIRS = 22860
 PROB(ATTRITION = YES) > PROB(ATTRITION = NO)
• P(1) > P(0)

18. SOMER’S D
• % OF CONCORDANCE - % OF DISCORDANCE
• FOR TRAIN
• SOMER’S D = 69%
• FOR TEST
• SOMER’S D = 63%

19. ROC CURVE
TRAIN TEST

20. RANK ORDERING AND KS STAT
Decile
Attrition-
No/0
Attrition-Yes/1 Total % of No % of Yes Cum % of No Cum % of yes Diff
(0.479,0.957] 31 72 103 4% 41% 4% 41% 37%
(0.297,0.479] 60 43 103 7% 24% 11% 65% 54%
(0.187,0.297] 85 18 103 10% 10% 21% 75% 54%
(0.124,0.187] 89 14 103 10% 8% 31% 83% 52%
(0.0864,0.124] 94 8 102 11% 5% 42% 88% 45%
(0.0586,0.0864] 96 7 103 11% 4% 53% 92% 38%
(0.0372,0.0586] 96 7 103 11% 4% 65% 95% 31%
(0.0201,0.0372] 100 3 103 12% 2% 76% 97% 21%
(0.01,0.0201] 101 2 103 12% 1% 88% 98% 10%
[0.0011,0.01] 100 3 103 12% 2% 100% 100% 0%
0%
20%
40%
60%
80%
100%
120%
1 2 3 4 5 6 7 8 9 10
Cum % of No
Cum % of yes

21. OBSERVATION
0%
2%
4%
6%
8%
10%
12%
% contribution
% contribution
 Single male
 Low job level employees in the age
group of 18-25
 Sales Representatives who are
Frequent traveller
 High experience with low salary
 As distance from home reduces work
life balance

22. RECOMMENDATIONS
Flexible working hours(OT)
Promote team building activities(ES)
Training, Job Rotation(JS)
Gym/ Yoga, Crèche facility , Medical Insurance (WLB)
 Video Conferencing(wherever possible),Travel on
Rotation Basis(TF)
Work From Home(DH)
Mentoring of New Hires, Promotion(JI)
Keeping Achievable Targets, Job Promotion(SR)