My Guided Project #3: Exploring eBay Car Sales Data

I would like to share my third Guided Project to get feedback from the community.

I spend over a month on this project, I think. While working, I discovered that Python methods such as ‘.groupby()’ provide a lot of options that should make data analysis much easier and faster, so I didn’t try to write complex code, but rather dive into the coding logic
Also, I discovered that there is a ‘.corr()’ method for calculating correlations, so I’m going to look into that as well instead of writing long lines of code.
And since I’m sure in fact, I would use any mechanism for rendering and visualisation, for example, Preset\Superset, I didn’t try to write long code for charts either.

However, I tried my best and really learned a lot during this project.
I hope to get a few words of criticism to improve my path in Python).

Exploring_eBay_Car_Sales_Cleaned.ipynb (566.2 KB)

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

1 Like

Hi! Thank you for sharing your project!

I didn’t have much time to look into it, but I really liked that you were using quite a lot of things other than what you have learned so far here on Dataquest, which we’ll be necessary in a working environment as well. So you are on the right track :slight_smile:

Something I noticed is that you created a function to convert strings to datetime objects. If you check, you’ll see that your columns weren’t actually converted (they are still object types, rather than datetime64). Unfortunately, I couldn’t figure out why your function doesn’t work, but pandas actually has a function pd.to_datetime() which converts strings to datetime type easily.

Thanks for your feedback! I really put more effort into using whatever I knew rather than the following instructions. I can even say that I was surprised when I got to the last pages of the lesson and found that I had already done by myself what we are asking.

About these data time objects. I also noticed this strange fact. Of course, I tried to understand and fix it. Then I had a situation where Pandas and Numpy consider these objects to be not strings, but data-time objects. So, I left them as such, since I could operate on them as data-time objects.

About the pd.to_datetime() function, thanks. Python is really something, it provides functions, methods and little tricks to get the result in a nutshell instead of writing long complicated queries. It is the genius invention.

1 Like

It’s difficult to deal with datetime objects, isn’t it? I’m redoing this guided project at the moment and I’m really struggling with it :sweat_smile: