I would love to be able to select a period on the graph and see the min, max, average and for periods where it inclines/declines, how fast it did so (derivative). That way I can get insight in how fast a room heats up/cools down.
Exactly the kind of thing I’d like to do as well. My mindset comes from decades at HP/Agilent/Keysight where the business was built around real-time acquisition of data and then slicing/dicing it a bazillion different ways to help figure out what happened and why. 85% of the UI for a scope is just this. I’ve resorted at times to dumping the HA data, reading it into the scope SW as a waveform, and then doing what it does best.
If there’s FOSS that comes anywhere close to this then I’ve not found it. If someone has - PLEASE POST A LINK.
Just today I wanted to create an automation that would trigger on the output of a derivative helper given an input just like the history above. I wanted to look at the hot water heater’s response to water demand to decide when I should turn on a boiler fed HW preheater to help out. I created one but saw it only worked on “new” data, so I’d have to run long experiments to get enough data to figure out automation triggers. I have lots of great historical data but I found no way to get the helper to use it. So I have to think up Plan B.
There are lots of potential "Plan B"s (e.g. Grafana, pgsql queries, …) But it should be a lot more straightforward.
Just to tickle a build on the original idea, “measurement based modeling” was a big use for instrumentation. A quick look at the waveform above shows at least one, perhaps more, exponential time constants, with bumps, wiggles, and noise superimposed. A model for the room that includes response to heat input, loss to the outside, maybe doors, lights, or comings and goings, can be inferred from the data and refined with simulations. Not an everyday necessity, but that’s in fact what I’m starting to do now that I’ve instrumented my hydronic heating system and want to figure out what to do to optimize it.
Other stuff like de-embedding (deconvolution), etc., start to make sense once the basics are available. At least for me, the big return on HA is being able to see what’s happening in my home and make smart decisions about how to improve it.