Medical Insurance Portfolio Project (Feedback wanted!)

[tonyponcedeleon22016]

I wanted to share my first ever portfolio project and hear some feedback if anyone has time. I didn’t do anything to crazy like regressions or predictions, but I just decided to break down the data by region and analyze that way. What I found was some variable mattered by regions, and some kind of didn’t. And it left me with a conclusion, but with more questions that with more data by region, I could probably analyze further. So it’s very much scratching the surface. Any feedback on style or the best way to implement classes here would be appreciated (classes are a weak point in my python skills!) But I was proud of myself for being able to undo the csv in to lists and easy to use dictionaries.

tonypdl/med_insurance_project

In my repository, I have a few files, and the blank jupyter file I didn’t use. I ended up just commenting in my script.py as I went. I’ll need to get more comfortable with jupyter!

[lisalisaj]

Congrats on finishing the project!
It looks like you have a good grasp of the concepts from the lessons.

I will say that it might be better practice to not have all your code in one cell. (if you get an error, then it’s a pain to determine where exactly in the code the error lies). Also, you can’t see the output of the cells, nor can anyone who looks at the file. So it’s a little difficult to follow what you’re doing and your thought processes.

More to that point, data scientists and data analysts are storytellers; we are conveying our research findings to a technical or non-technical audience. And, depending on the audience, you can decide how technical to make your presentation.

For example, you could add a text cell before a code cell that explains what you’re doing/what your thought process is for writing that function or whatever, or maybe have some comments in the code cell itself. You already have comments in there, but again, we can’t see the output in the notebook.
When you write that function about smokers and regions, it would be helpful to see the visual output of that function within the notebook rather than looking at a .png file.
Even though it’s a jupyter notebook and not a presentation in Slides or PowerPoint, it still has to flow like a presentation…people have to know what they’re looking at and why you found it important to investigate.
Also, it could be helpful to add a READme.md file to that repository—so you can describe the data set, where it came from etc.

Here’s an example from a project that I did about GDP and Life Expectancy at Birth:

Just a one sentence explanation about the plot, and then I discuss it further afterward.

I hope that makes sense?
Happy Coding! And thank you for posting your work!

[lisalisaj]

Also, one line of code that you can add to the top of your notebooks after you import matplotlib is:
%matplotlib inline

You only need to run that cell once (unless your kernel session times out, then you have to rerun your cell w/all your imports).

which embeds the .png graphic after the cell.
More on that:
https://jakevdp.github.io/PythonDataScienceHandbook/04.00-introduction-to-matplotlib.html