Guided Project: Creating a Kaggle Workflow - Titanic

Any comment is welcomed :slight_smile:

Creating a Kaggle Workflow.ipynb (116.6 KB)

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

1 Like

Hi @JamesDiu,

Thanks for sharing your project. You’ve done a good job! I especially liked your very neat visualizations, well-structured and easy-to-follow code, and the way you defined the majority of your functions, with all the necessary descriptions and comments (and in general, in was a great idea to introduce all those functions for this project).

Here are some suggestions for your consideration:

  • There are some warnings around the project. You can think of how to fix them, or, alternatively, just suppress them :slightly_smiling_face: This code would be useful for this purpose:
import warnings
warnings.filterwarnings('ignore')
  • I would suggest to you to add a more expanded introduction to the project, including the link to the dataset you used, and also a conclusion summarizing the obtained insights.
  • A good practice is to re-run the already ready project, to have all the code cells in order and starting from 1.
  • Also, a good idea is to use a uniform style of the quote marks for the string data type all around the project: or only single, or only double.
  • The code cell [8]: here you can consider applying a for-loop for despining the plots (for all ax.spines)
  • It’s better to unite the subsequent code cells if they don’t have any output or any markdown comments directly after them. For example, [1]-[2], [3]-[5], [8]-[9], and some others.
  • The code cells [9]-[11] (well, actually it derives from the code cell [8]). Here it’s better to remove the legend (since it consists only of one item), and also not to rotate x-ticklabels, since they are just short numbers.

Hope my ideas were useful.
Good luck with your future projects!