Clean And Analyze Employee Exit Surveys - Raghav_A

Hello All!

Feel free to go through and provide me with feedback on my Guided Project, so that I can improve my work further!

Thanks!
Raghav_A

https://app.dataquest.io/m/348/guided-project%3A-clean-and-analyze-employee-exit-surveys/11/next-steps

New GP6.ipynb (342.9 KB)

Click here to view the jupyter notebook file in a new tab

5 Likes

Hello @raghavaseem! Thanks for sharing.

I liked how you documented your project and explained each step/decision you make.

Here are some ovservations:

  • You can add some comments to your code, especially to functions, even if they seem self-explainable. You will thank yourself in some months if you ever get back to the project.
  • In the function category_of_service you can use the OR operator to avoid repeating the New or Veteran categories.
  • You should add title and axes label to your plots to make them clearer.

You can find more project’s style tips here.

Happy coding!

3 Likes

@ artur.sannikov96

Thanks for the feedback! Your inputs have been duly noted, and I’m pretty sure they will help me going ahead.

Cheers!

2 Likes

I like your presentation of the data preparation and analysis. It’s helped me decide to put more of my markdown commentary before each code section, as you have done, rather than after, as encouraged in the Dataquest instructions. I also like your use of python comments where breaking things up into separate code and markdown cells seems too “heavy.”

In cells 40, 46 &46 you show the service categories sorted in alphabetical order. Unfortunately this obscures the ordinal relationship between the categories. In my project I came up with one solution to this in cell #11 in my analysis notebook.

Be careful about the language you use and the impressions they leave.

For example:

Permanent Employees are more dissatisfied than Temporary/Casual/Contractual workers

We can’t conclude this, because this dataset only covers employees who have left their positions, rather than all employees, and we’ve focused only on those who resigned.

Further, keep in mind that there are two sides to any binary comparison. If people aren’t resigning due to job dissatisfaction, why are they resigning? What other reasons do Temporary/Casual/Contractual have for resigning? If you think about it, those reasons may relate to the reasons they weren’t permanent employees in the first place. They might prefer a permanent position, somewhere, and resign when offered one. They might have other priorities, like school, or family, or another job that matters more to them, and, given a conflict, they might resign in favor of the higher priority. It could be that Temporary/Casual/Contractual are just as satisfied or dissatisfied as permanent employees, but the % of them citing job dissatisfaction is lower because more of them have other reasons.

Finally, keep in mind the big picture. This dataset doesn’t tell us about overall employment in each institute, each category, each age group, etc. What if there are 10x as many teachers as there are administrators? If that’s the case, then teachers look, overall, more satisfied than administrators.

I’ve struggled a bit with communicating these things clearly in my own projects, so I hope what I’m saying makes sense here.

3 Likes