Project: I-94 Traffic Indicators

Hello Community! Here is another project that I am happy to share with you all.
Please provide your suggestions and feedback. Happy and eager to learn from you all!

3 Likes

@sahiba.kaur.stats congrats on completing this project :handshake:. You’ve done a great job :rocket:

I really liked how you analysed traffic on holidays and based on temperature. I’ve not seen that before, and I haven’t done that myself, so kudos.

I have a couple of pointers that I hope will help improve the project. (please click on the triangle bullet points below)

Presentation Style
  • I recommend the inclusion of an introduction section with details that include the context behind the dataset, a link to the dataset and an idea about the source of the dataset. This allows readers to go over the dataset on their own.
  • I noticed that you started identifying and removing outliers based on temperature, rain, cloud etc. It is the first time I have seen that. If you did a correlation for these columns and traffic volume, you would notice that it was insignificant. I would have, therefore, avoided this. It would be interesting to see what opinions others hold on this regard.
  • It would be really helpful to readers if you could highlight the number of rows as you removed your outliers.
  • I really like this bit of code. I wasn’t aware that you could do this:
a=data.groupby(["year","month"])["traffic_volume"].mean().reset_index()
a.set_index(["year","month"],inplace=True)
a.unstack(level=0).plot(figsize=(12,10),subplots=True,layout=(4,2),legend=True)
plt.show()

I was however confused about why for some years the plots did not show the full data. e.g., 2014, 2015 etc. I think it would be good to explain this data loss

  • I noticed that you pointed out the Philando Castille shooting. It would be helpful to readers if you could provide a link to news articles containing details of the same. Specifically, something you went through to get the details.
  • For the output of cell[21] I would suggest switching axes so that the days of the week come on the Y axis and the traffic volume comes up on the X-axis. It makes it easier to read.
  • I recommend the inclusion of titles to plots to improve their readability.
Coding Style
  • Most of your plots have been created using dataframe objects. I recommend that you dive deeper in to Matplotlib libraries and use the same for plotting as they are more flexible and have more parameters you can manipulate
Bugs/Inaccuracies
  • Check out this project by @Elena_Kosourova with analysis of holidays. I think there are a couple of outliers in there that you might have to sort out before you can do a full analysis.
Miscellaneous
  • Nothing much to add here

Overall a great job :100: and keep the spirit :fire: going.

1 Like

Thank you @jesmaxavier for the detailed feedback. I appreciate the time and effort you took to go through my project.

  1. Now that I have given the project another look, I completely agree that the removal of outliers was not really necessary for the project.
  2. For some years, the plot did not show any data because it was not available for some months.
  3. The ambiguity in the “holiday” column was something that I completely missed. Kudos to @Elena_Kosourova for making such a detailed observation on it.

I would like to thank you once again @jesmaxavier for the suggestions.

1 Like