Let’s assume I have quite a few sensors that report numeric values. I start using these sensors in graphs, automations, etc. Then I realize that the sensors can produce outliers. So I use the outlier filter. However, this creates new entities and I have to adjust all graphs, automations, etc. Next I realize that some values have to be shifted since there is a systematic error as well. Therefore I create template sensors which shift the values. And again I have to adjust everything since the entities have changed again. And it continues…
I was wondering if there is an approach to make sure that “end-users” of sensors in HA don’t have to be touched every time I add some transformation between the raw value and the final one. Of course I could create dummy template sensors which simply “rename” one of the intermediate sensors to the final value but this doesn’t scale very well for dozens or even hundreds of sensors.