🔥 It looks like it's time to lock in the current version of the machine learning model (LULC v.6.0). Validation results on an independent test dataset (Confusion Matrix) demonstrate the excellent predictive power of the algorithm: 🔹 Overall Accuracy: 98.29% (95% CI: 98.16% – 98.41%) with a No Information Rate = 60.28% (p-value < 2.2e-16). 🔹 Cohen’s Kappa: 0.9691, confirming high classification reliability even with severe class imbalance. The model shows a very low error rate for non-vegetated areas (Solid, Balanced Accuracy 99.72%) and water bodies (Water, 99.75%), but expectedly faces challenges at the boundaries between the Lawn (open grass) and Park (sparse trees) classes. Currently, I am moving on to an advanced analysis of the drivers shaping Urban Heat Islands (UHI) in Calgary and preparing a publication for my website. Stay tuned for updates! #Calgary #OpenData #UrbanHeat #DataScience #ClimateResilience #YYC #Geoscience #CityPlanning #RemoteSensing #RStats #MachineLearning #GreennessOfCalgary