Employee Dissatisfaction More Evident in Resignations After Longer Employment Terms

Hi Everyone,

Here is a basic analysis of the Employee Exit Survey data.

I check the output at each step because it’s so easy to misapply a cleaning step imho.

I thought it was important to actually order the bars of the final plot by length of service or else the story is not clear. Unfortunately I didn’t actually execute that perfectly but you get the idea.

The project description can be found here.

Feel free to download my notebook: 2.4 Guided Project Resignation Dissatisfaction.ipynb (238.8 KB)

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

1 Like

Hello @kwu! Thanks for sharing your project:)

I liked that you explained well every step you’ve done and you also find a more elegant way to convert the Job Dissatisfaction to True or False in the TEFE. Also good job on proposing some improvements to the project:)

What can be improved:

  • Reduce the number of sections (for example, the first two sections (Data Set - At A Glance and DETE Survey - At a Glance can be merged)
  • Do not reference the DQ instrtuctuion. These are just guidelines and the project is yours. Explain why you drop column (why they are irrelevant), motivate your code with your own thoughts!
  • Do not include too many technical details in Markdown (like in the section DETE Survey - At a Glance)
  • You have some typos and also unfinished sentences (like A quick check at Established column data indicates institute_serivce durations of 7,8,9 and 10 years were included and they were.)
  • Be aware of the SettingWithCopyWarning. It may cause problems. You can read more about it here
  • Improve you plots: label the axes, title them, remove top and right spines, remove the legend. The title you’ve given to the plot is more a subtitle and it good but do not forget about a shorter title:)

Happy coding :grinning:

Hi @artur.sannikov96,

Thank you very much for reviewing my project and taking the time to provide all this feedback.

I have been struggling with the ‘voice’ to be used in writing the projects. I have been documenting my learning process as a way to reflect the analysis thought process but I get a little confused by my target audience - who are fellow learners like myself but also I would like to be building a portfolio to share outside of the dataquest community. I feel you are nudging me to focus more on the latter, would that be accurate?

I have also been working on finding the right amount of explanation of my code. For this I feel my audience in the real world would be more myself and anyone who might want to modify my code so that who I am trying to target.

I definitely intend to focus energy on improving my visual plots, especially now that I feel more confident with generating the content. The plots are the most exciting part!!

I did not have a chance to incorporate all of your feedback into my recent guided project, but if you are interested to see my most recent work, it’s just been uploaded here.

Many thanks again for your feedback, it helped clarify some aspects of the guided projects I was uncertain about.

Hello @kwu! I will try to review your next project in the next few days;)

I stressed the fact of building a portfolio and have your possible employers as your target audience because at some point you will have to create independent projects. DQ GPs are good but they are guided which is usually does not happen in the real world. So if you can train yourself to target not only fellow learners but also other categories for the future career it will help you when you’ll makeindependent projects:)

Happy coding:)

Hi @artur.sannikov96 ,

Thanks very much again for the guidance. For the recently completed project I referred you to - because of something unexpected coming up, its target audience really is the community and not a potential employer.

So, rather than checking that out, could you instead please provide a bit more detail on how you would recommend approaching the guided projects? It seems to me as though the guided portion ideally should be completed as instructed, but for the ‘next steps’ I should continue with my own analysis by removing any code irrelevant to my own project goal (or start a fresh notebook). So in the end I have 2 notebooks (in jupyter) - one for the guided portion and one that is my stand-alone presentation to to be shared with the community AND eventually potential employers.

I apologize if this is really obvious but thanks for your patience to help along my DS learning path!!


Hi @kwu!

You could do the next steps in the same notebook. And if you that some guided steps can be improved or done in another (potentially better) way, go for it! This is the moment when you start learning:)