Posts

Showing posts from February, 2020

Sampling methods for visual inspection or quality control of GIS data

Common quality errors for visual inspection 1. Missing features 2. Redundant feature 3. Misplaced 4. Wrong shape (line and area) 5. Miscoded (wrong attributes) Typical process for visual inspection Determine sampling method Create a sample Inspect sampling record Mark as pass or fail Generate report Determine acceptability Common sampling methods  By Fixed number - randomly choose features based on a fixed number of sample that can be carried out by the team. By grid or polygon - create an index grid and sample features per grid.  By percentage of features - determine the sampling percentage of all features. Amount of features for inspection/QC is based on direct percentage count or applied weights, if there is any. By calculation - determine the statistically significant number of samples at a certain confidence interval, plus or minus the acceptable error. ( Given the sample size, how many features can fail inspection before my entire database fails.