I got some KNX temperature sensors in the house and I’d like to use the long term statistics of HA.
I have configured one of the sensors like this:
- name: "Büro Temperatur Taster Türe"
state_address: "4/1/2"
type: temperature
state_class: measurement
This works nicely and I can use the statistics graph card but I can’t believe that the mean value is really correct.
This is how it looks for the last 90 days. So max/min/mean look nearly the same. Is that really correct?
Those statistics are on a 1h basis, so it makes the min/mean/max temperature over 1h don’t vary that much.
Hmm ok. So these charts don’t make much sense. It would be great if the mean value would be configurable.
Not sure what your expectations are, but so far, for real statistics, InfluxDB is the way to go.
Well the mean value in 1h doesn’t make sense since it follows the other values especially for room temperatures. I’m using InfluxDB plus Grafana but a more useful integrated solution would have been nice.
Long-tern statistics are pre-calculated.
Not sure why you think 1h makes less sense than anything else for temperature, tbh, especially for indoor ones with limited variation. Generally speaking, it’s nice to have but doesn’t tell much.
Well what is the sense of these three graphs? I don’t see much of a benefit in the visual representation. The long term stats are nice but I don’t get the sense of these max/min/mean.