DATA SCIENCE | Urban Delivery Demand Prediction | Final Progress

[liajohn30]

Reason:
Final pull request for notebook consolidation, model evaluation, reproducibility improvements, and preparation of final report content for the Urban Delivery Demand Prediction project.

Changes made:

• Finalised and reorganised project notebook
• Refined T-GCN implementation and evaluation workflow
• Added model comparison observations and documentation
• Generated overall and zone-wise evaluation outputs
• Improved notebook readability and reproducibility
• Added final experiment outputs and saved model artifacts
• Prepared analysis content for final report and presentation

[molliefernandez-mentor]

@liajohn30 your pull request does not have any reviewers, please assign a reviewer or close the PR

[PriyanshuCauleechurn]

Reviewed the notebook and overall the workflow is well organised and easy to follow. The markdown explanations help provide context throughout the notebook and the progression from preprocessing to modelling is clear.

A few areas that could be strengthened are adding slightly more interpretation of the EDA visualisations and expanding on how key findings influenced later modelling decisions. Some of the later experimental sections could also benefit from a short summary of the main takeaway or outcome.

I also noticed that the PR currently contains a large number of files (29 files in total), including multiple CSV files. It may be worth reviewing which files are necessary for the final submission and removing any datasets or generated files that do not need to be stored in the repository. This will help keep the repository cleaner and align with the expected project structure.

Overall, the notebook is well structured and technically strong, with only minor improvements needed to further enhance readability, documentation, and repository organisation.

Happy to approve but do check in on the number of files.

[Litxinh123]

Hi Lia,

Good work overall. I have a few suggestions to make the submission clearer and more consistent:

• Please move all CSV dataset files under /DEPENDENCIES folder under the /datascience directory instead of keeping them scattered across this working folder
• Please make sure all three required files are included: .ipynb, .json, .html
• In the notebook .ipynb, please follow the required template in the introduction section. You can refer to the example screenshot below.

Let me know if you have any questions or concerns.

Thanks.