If the number of dataset saw a massive jump (50%), then back down a week later, that seems more like the correction of an error.
If the number of dataset saw a massive jump (50%), then back down a week later, that seems more like the correction of an error.