DATA SCIENCE | Project 22: Parking Utilisation and Management | 100% Completion | READY TO PUBLISH

[AtishayDeakin]

This is the final pull request for Project 22, representing the complete technical handover of the predictive parking asset to the Melbourne Open Playground (MOP) repository.

List of Changes:

1. End-to-End Pipeline Finalisation: Successfully processed 14.2 million rows of sensor data with full ISO 8601 standardisation.
2. Strategic Anomaly Management: Implemented robust filtering logic to handle COVID-19 pandemic outliers, ensuring model reliability during varying urban mobility phases.
3. Advanced Feature Engineering: Executed a strategic pivot from basic time-series features to high-granularity Bay-level encoding and Lag Feature implementation (prev_occupancy).
4. Model Optimisation: Accomplished a significant performance jump, moving from a 49% baseline to a final validated 84% accuracy using an optimised Random Forest classifier.
5. Visualisation & Interpretation: Integrated comprehensive spatial and temporal visualisations pinpointing peak demand hotspots on Lonsdale and Bourke Streets.
6. Documentation: Finalised the Jupyter Notebook in a step-by-step tutorial format for seamless handover to Chameleon.