Closed kerhbal closed 2 years ago
but it's a fixed one
Try lower_bound: ~40
My thermostat sometimes just return a -124% humidity
These wrong values may be filtered by using a template sensor.
Some my devices return a 99999
value - I have to filter it in a template sensor.
@ildar170975 I see. since lower_bound: ~10
will allow any data to break it as shown below(use 10 as example here):
Though I haven't tried the template, should be a good solution. thanks!
Me too - for humidity graphs I use ~10
since in my apartment it is usually 25..50%.
Here is a good explanation of the "soft bound":
https://community.home-assistant.io/t/lovelace-mini-graph-card/71385/1372?u=ildar_gabdullin
The y-axis is normally set to the values of lower and upper bounds, but if the graph value is lower than the lower bound or higher than the upper bound, the y-axis is expanded
.
Hope this is a correct description, it usually matches my observations.
@kerhbal Please close the issue if it is resolved
@ildar170975 yeah sure will try it later today. thanks!
the filter template works! thanks!
the filter template works! thanks!
can you give me a link or a temperature? I meet the same question.
@kettly1260 I have something like this in configuration.yaml to create a filter based on my humidity sensor
sensor:
- platform: filter
name: "filtered_humidity"
entity_id: sensor.humidity
filters:
- filter: outlier
window_size: 4
radius: 40
I don't quite remember what the numbers mean... you can check it here https://www.home-assistant.io/integrations/filter/
and this filter acts as a new entity(a virtual one), then you can ignore the old entity and just put this new entity in any graph/frontend
hi, first of all this is really fantastic, thank you for developing it! Background My thermostat sometimes just return a -124% humidity...like this:
I know there is lower_bound we can use to limit the graph, but it's a fixed one. If I choose to set the lower bound to 40, I cannot see the ones below it, but if I set it to be 20, the graph will be more flat and makes less sense. What would be best If we can ignore all the data points below 0 before generating the graph, that will be best. In this scenario, we can assume no data points in the period. Hope this makes sense, thanks!