Data Analytics for Digital Marketing Lecture for Advanced Digital & Social Me...
Mohammad Shaiz Akhtar-POC.pptx
1. TAKENMIND GLOBAL DATA
ANALYTICS INTERNSHIP
PROOF –OF- CONCEPT
NAME: MOHAMMAD SHAIZ AKHTAR
COUNTRY: INDIA
LINKEDIN: HTTPS://WWW.LINKEDIN.COM/IN/MOHAMMAD-SHAIZ-AKHTAR-5439B5169/
EMAIL: SHAIZMOHAMMAD@GMAIL.COM
PROJECT TITLE: ATTRITION CONTROL
2. PROBLEM STATEMENT
The data is for company X which is trying to control attrition. There are
two sets of data: “Existing employees” and “Employees who have left”.
OBJECTIVE:
1. What type of employees are leaving?
2. Which employees are prone to leave next?
3. APPROACH
Read the given excel sheet using pandas which has an in-built function to
read (.xlsx) files.
Preprocess the data to check whether any of the worksheets have ‘NaN’
values.
Once data is preprocessed, we will perform univariate, bi-variate and
multi-variate analysis to understand the behavior of features by plotting
on plots and graphs.
4. Once we read the excel sheet, we find out that there are two worksheets present in the excel file, namely- ‘Existing’ and ‘Left’ employees of a company.
There are 10 columns in each table. They are:
1. EmpID
2. Satisfaction_level
3. Last_evaluation
4. Number_project
5. Average_monthly_hours
6. Time_spend_company
7. Work_accident
8. Promotion_last_5years
9. Dept
10. Salary
Major features affecting an employee’s attrition:
1. Satisfaction
2. Promotion
3. Average Monthly Hours
4. Salary
5. UNIVARIATE ANALYSIS
Satistisfaction
1. Employees who left: 40% satisfaction
2. Existing employees: 70% satisfaction
Average Monthly Hours
1. Employees who left: 225 hrs
2. Existing employees: 200 hrs
Salary
1. Employees who left: ‘Low’ salary employees domination is observed.
2. Existing employees: Roughly equal number of ‘Medium’ and ‘Low’ salary
employees is observed.
7. BIVARIATE ANALYSIS
From the Box plot of both Existing Employees and those who Left, in terms
of number_project we can discover the following insights:-
Employees who left - The median Satisfaction level of employees who
have done 5 projects is around 60% and after 4 projects satisfaction level
decreases.
Existing employees - The median Satisfaction level of employees who
have done 4 projects is around 65% and after 4 projects satisfaction level
decreases.
9. MULTI-VARIATE ANALYSIS
Satisfaction Level
Employees who left - Satisfaction level of the employees who were not
promoted in the last 5 years was around 40% in all the departments
whereas satisfaction level varied for those employees who were promoted
in the last 5 years, also the employees were promoted in only few
departments(very few in sales, IT, management and negligible in technical,
support, management)
Existing employees - Satisfaction level of the employees who are not
promoted in the last 5 years is around 70% in all the departments whereas
satisfaction level varies(but doesn't fall below 60%) for those employees
who are promoted in the last 5 years, also the employees are promoted in
all departments excluding IT and product management
11. PEOPLE WHO ARE PRONE TO LEAVE
People having low satisfaction are more likely to leave.
Higher average monthly working hours directly affects the attrition rate
of an employee causing employees to leave.
If the number of project exceeds 3, the employees of the company are
more likely to leave.
If the salary is low and the number of projects assigned to an employee
is larger, employee is more prone to leave.
If average monthly hours is greater than 200 hrs and time spent is
more than 4 years, then there is a possibility of the employee to leave.
12. CONCLUSION
In other to prevent future attrition and constantly maintain a high
satisfaction level of employees, the company should ensure the number of
projects allocated to an employee is not too overwhelming thereby losing
enthusiasm which may result to seeing a need to leave
13. Thank you TAKENMIND for this Internship for helping me improve my
Data Analytics Skills