DATA SCIENCE | Project 22: Parking Utilisation and Management | 50% Completion

[AtishayDeakin]

This PR merges the completed Sprint 1 (Data Setup) and the initial phase of Sprint 2 (Model Build). It establishes the end-to-end data pipeline, from raw ingestion to a functional baseline predictive model.

List of Changes:

1. Data Standardisation: Implemented a Python-based ingestion pipeline that converts raw sensor pings into a consistent ISO 8601 format.
2. Occupancy Mapping: Developed logic to map categorical sensor states into a binary format (1/0) for machine learning readiness.
3. Strategic Filtering: Added a custom filter to handle COVID-19 lockdown anomalies (March–May 2020) to maintain baseline model integrity.
4. Exploratory Data Visualisation: Created time-series and spatial plots to identify high-congestion hotspots in the Melbourne CBD (Lonsdale and Bourke Streets).
5. Baseline Model Implementation: Developed and trained a Random Forest classifier to forecast bay availability.

[manya0033]

Hey Atishay, I have gone through the notebook and the PR. Good progress on the pipeline and baseline model, the logic is solid. However, there are a few things that need to be addressed before I can approve, based on the PR checklist:

1. Dataset access - The notebook loads from a local CSV. As Per the checklist, datasets need to be accessed via API v2.1 and API keys must not be visible.

2. File naming- The notebook is currently called task.ipynb. It needs to follow the proper naming convention.

3. Australian English - There are a few American English spellings that need updating:

   • "Standardizing" - "Standardising"
   • "Visualizing" - "Visualising"

4. Libraries not at the top - matplotlib, seaborn, and sklearn are imported halfway through the notebook. The checklist asks for all library imports at the top.

5. No markdown cells -The notebook is 4 code cells with no markdown. The checklist says the use case should read as a clear step-by-step tutorial using the correct Use Case Template. Adding markdown sections to explain what each step does and why would bring it in line.

6. Visualisation interpretation - The occupancy plot has a title and labels which is good, but the checklist also requires a written interpretation of what the visualisation shows. Adding a markdown cell below the chart explaining the key takeaways would cover this.

7. Citation artefact - Cell 1 has [cite: 317, 349] in a code comment which looks like a leftover and should be removed.

8. Completion level - The PR title says 50% but the notebook is currently 4 fairly short code cells with minimal elaboration. For 50% I’d expect to see more depth in each section - data quality checks, result observations, more detailed EDA with visualisations, and discussion of model performance beyond just the accuracy score.
   Once these are sorted, I'm happy to re-review. Tag me when you push the updates.