Exploratory analysis of training samples for landcover classification of an area in northern Alberta. Following on from previous data explorations, standard deviations of objects and certain topographic indices have been discarded. Final variable selected was based on avoiding correlated variables.

Exploratory analysis of training samples for landcover classification of an area in northern Alberta. Sampled as objects (clusters) with Google Earth Engine's SNIC algorithm. Means and standard deviations of objects were examined, and differences between seasons.

Spectral profiles, density plots, box plots, and correlation matrices of training data (pixel based) against S1, S2, and topo derived variables. Code adapted from Kamusoko (2021) Explainable Machine Learning for Land Cover Classification: An Introductory Guide..https://aigeolabs.com/ebooks/

Violin plots of training data (pixel based) class against S1, S2, and topo variables - Oct 3021

Analysis of variables from training samples used in the Rocky Mountain Alberta ecoregion wetland classification. Variables are derived from 2020 Sentinel-2 optical data and ALOS DEM derived topographic indices.

Compares two open source tree canopy height datasets for the province of Alberta, Canada with our organization's data. Density plots, box plots, line graphs, bar graphs made made using ggplot2. Alberta Biodiversity Monitoring Institute.