School project · EE4211
Carpark Availability Prediction
Applied ML models to 5 months of real carpark data to predict availability near hawker centres — visualised as an interactive heatmap.
The problem
Drivers near busy hawker centres often circle for parking with no visibility into which nearby carparks are likely to be full at a given time of day.
What I did
Part of a 4-person team. Collected live carpark data via public APIs over 5 months, cleaned and structured it for model training, compared multiple ML approaches, and built a heatmap visualisation for carparks within 200m of hawker centres in Northern Singapore.
Outcome
Demonstrated that historical occupancy patterns are strong predictors of future availability. The heatmap proof-of-concept made the results immediately interpretable.