As a data scientist or analyst, sometimes you need to restrict access to models or data Viz dashboards / MI you suspect are no longer used or useful to meeting business needs.
I periodically do this too - "turn-off" reports / dashboards and wait to see who squeals.
1/3
You will always learn something from the responses or lack thereof.
If people aren't using it, there's a reason. Maybe you need to make the MI more relevant to a role or process that has evolved (concept drift), maybe you're getting poor or inaccurate results from a changing process / data input schema (data drift), etc.
Or maybe it's simply time to hit SHIFT+DEL.
Every well done dashboard design and deployment life-cycle should also have an 'end-of-life' or 'model retirement' plan
2/3
Do you ever turn off access to your models / MI and dashboards to see if anyone even notices? 🚫✅
@paul_ekwereii I find metrics/reporting on the usage of the analysis to be more useful. Who is using it? When are they using it? Why are they using it?
Should they be using something else instead? Are they using the dashboard/analysis correctly? Do they understand the assumptions/limitations of the dashboard/available?
Turning it off is just a boolean; the rest is more interesting.
ai-Paul E
Don Armstrong