Guided Project: Exploring Ebay Car Sales Data - First Project Post!

Hi all!

I’ve been working with Python/pandas for a bit now and want to start doing more polished projects to add to my portfolio. This is the first project that I’ve tried to make “readable” and presentable for other’s, so please let me know if there are any areas where you think I could improve my readability and ease of understandability for an audience.

I also added a few “extras” that weren’t in the guided project instructions (i.e. I lightly used matplotlib) since it helped me to better organize my thoughts, but the general conclusion was still in line with was was asked in the guided project.

Basics.ipynb (203.3 KB)

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

2 Likes

Hi @anna.strahl,

Great job completing the guided project.

This is a great mindset to have since there’s a great emphasis on communication and storytelling in data science.

Feel free to add more “extras” if you want. The guided project is a starting point and the instructions are mainly for teaching and reinforcing previously learned topics. And if you want to include the project in your portfolio, it can be beneficial to deviate from the instructions (when appropriate) and add some “extras” to make the project truly yours and not just a project given by Dataquest.

Moving on to my thoughts on your project:

  1. An introduction is missing, or specifically, it appears to be because the project title is missing. It also seems to be missing because the project abruptly starts with a code cell.
  2. Move the %matplotlib inline line from cell [1] to [2].
  3. The overview of the variables is a great step towards helping the readers to learn more about the data set. Nice work.
  4. For the breakdown of the variables, consider using a Markdown table to make it more presentable. I like to use this table generator and complement that with learning how to align table.
  5. This depends a lot on who your readers would be but in all the text sections in your notebook, try to use as little technical jargon as you can. Examples: change"int/float" to “numbers”, “string” to “text”, “null” to “missing”. A technical reader will probably understand what you meant, but a non-technical one will probably scratch their heads or be forced to look up things on the internet. It’s always a judgement call because using less technical terms has the tradeoff of losing some nuances that are captured by the more technical terms, and as I mentioned before, it all depends on who you expect to read the project. It’s also very challenging to make technical things accessible to non-technical people but it might be worth it.
  6. " from camelcase to snakecase" → this might need a simple example to show the difference between the two i.e. for the non-technical readers.
  7. If you’re interested to use matplotlib more, the outliers section can be a good spot to use boxplots.
  8. The histogram looks a bit odd probably because of the outliers and seemingly empty right tail. Try playing around with the bins value and see if it any of them can properly capture the range of values without any of the bins/bar becoming virtually invisible.
  9. The use of the ebay article to justify the cutoff point is quite creative. It never occurred to me that you can use an article like that to decide on an upper bound for the price of car sold on ebay. The only thing I’m not sure is if the data set uses German currency or US dollar; the article is appropriate if the currency used in the data set is the US dollar.
  10. The line charts need an x-label at least, to make it clear that it’s for dates. I’m also interested to know why you make the xticks empty.
  11. For the registration year line chart, in which you noted an anomaly in 2015, using matplotlib’s axvline (doc) is useful to pinpoint the anomaly in the chart. You might also want to add an extra tick for the year 2015.
  12. In the conclusion, you can also use the non-technical car brands name instead of using the the data set representation e.g. mercedes_benz. Try to wear the non-programmer hat for a moment and think “How would I write this if I’m not a programmer?” Intuitively, you’ll use the layman style such as BMW, Mercedes-Benz and Audi.

I find your project readable and I can follow your line of reasoning for each section. And I also think it’s great that you’re quite thoughtful when explaining your findings.

That’s all from me. Cheers.

2 Likes

Thank you so much for the constructive feedback! I’ve learned a lot from your comments and the resources you shared and have already implemented several of your suggestions in my newest guided project!

2 Likes

No worries @anna.strahl.

I’m glad my comments were helpful.

Thank you again for sharing your work with the community.

Keep up the good work. Cheers.

1 Like