TThe final step of the training module focuses on calculating the area covered by a target crop on a crop map. Simply counting pixels on a map to report changes is not accurate as maps contain errors that can lead to biased estimates. Systematic errors can over or underestimate areas. To address this, a sample-based approach is recommended to estimate areas and quantify uncertainty. However, when dealing with a stratification map, additional care is necessary as it may not correspond directly to the classes needed for area estimation. To obtain unbiased area estimates and uncertainties, you need to create a confusion matrix by comparing sample point labels from the algorithm agreement map to reference data. Then, use a spreadsheet and follow the equations and best practices from Olofsson et al. (2014 & 2020).