[APP][Pro] Insight Trends Reloaded

Hey,
I’m trying to get the time when someone took a shower. I have a humidity sensor in homey in my bathroom. I tried to use the “trend” check of ITR. But it doesn’t work. How can I calculate the trend between two datapoints, that I use the correct value for the ITR flow card.
Maybe I’m doing it the wrong way. But this is with what I came up with:

  1. If humidity becomes greater then 41-60%
  2. use ITR to check if the trend was greater than .75 in the past 15 min
  3. set var “showered” to yes/true and save current time in ms to var “showering time”
  4. for test purposes send a push notification to me

This is what the flow looks like? But it doesn’t work. Any idea what I’m doing wrong or am I going in a complete wrong direction?

This is what I can see in insights:


The Trend is the curve over time. Mostlikely not above 0.75.

Also I would change the trigger to something more simple.

As far as I know, I can use trend to identify rapid increase of a value or am I wrong?

I don’t like the trigger either, do you have a better idea?

Yes, but you would have to debug the trend value so you know how much it increases while showering.

I would take a trigger every X sec and if humidity is bigger then

I have the same flow. Or had it. I used when humidity is bigger then 65% then wait 2min if it still bigger then 65%. Then I set “showered” true

OK. I was way to high with 0.75. I now use 0.01. Let’s see if it works tomorrow.

It would look nicer, but from a performance perspective, wouldn’t that put more load to homey pro then my “solution”? Humidity is normally (apart from showering) not changing drastically throughout the day, so in my case if there is nothing to detect it might trigger the flow a few times a day, with every x sec it would be much more. :thinking:
Or is there another reason you would go for the x sec trigger?

Hi @L3nny, I do the same but instead of humidity, I use temperature. My flow looks like this and it works great except when the washing machine is on, but that is a general problem of the water trigger instead of the douche trigger.

To trigger the change I calculate the difference between “last value” and “first value” over a specific time. I think you can do the same to detect trend changes in your humidity values.

In raw data, you will see something like this.

And I chart.