Inventory accuracy and 'activity' in warehouses
This diagram simulates random samples from a population of 400. (State that it is prohibitively expensive to find inventory accuracy for all warehouses so a sample must be taken.)
Use it to show that random samples occasionally gives a false suggestion of an outlier, clusters or even a curved relationship, even though there are no outliers or clusters in the population.
Warehouses keep computer records of their inventory, but these are often inaccurate due to theft and paperwork errors. It was thought that the errors in these computer records may be related to the 'activity' of the warehouse — the volume of goods added and removed from the warehouse daily.
The data are artificial but are supposed to be the percentage error in the computer records of 400 warehouses and an 'activity score' for each.