What problem are you trying to solve?
When people are analyzing trends, everybody tries to redo the same work, explaining outliers.
Proposed approach
Use a shared library of known explaining factors for certain outliers and other odd events such as data errors. Stats.wikimedia.org already implements this. It might be a good example to follow, and to share their library, to some extent. https://meta.wikimedia.org/w/index.php?title=Config:Dashiki:Annotations/Wikistats/totalPageViews
What problem are you trying to solve?
When people are analyzing trends, everybody tries to redo the same work, explaining outliers.
Proposed approach
Use a shared library of known explaining factors for certain outliers and other odd events such as data errors. Stats.wikimedia.org already implements this. It might be a good example to follow, and to share their library, to some extent. https://meta.wikimedia.org/w/index.php?title=Config:Dashiki:Annotations/Wikistats/totalPageViews