I think the analysis graphs / resistence curves would be a bit more useful if we could also see the average indoor temp during the time considered. That’s because the “x degrees per hour” rate of change is sampled at the current house temp, so it’s important to show that, because if the house were much warmer or colder that curve would look different.
I’m imagining having several vertical lines: one for when each stage of equipment running and one for the resist curve (when no equipment is running). I’d expect the average temps to be close to the average setpoints during that time, but you never know…
Given the indoor temps are in a very narrow range, there’s no significant difference between any of them. All it ends up doing is splitting the available data points across different series and reducing overall accuracy.
It’s a good idea - just difficult to apply well given how many other factors there are.
Oh, I wasn’t suggesting splitting the data any further than it is currently, simply showing the average indoor temp for each dataset that’s being plotted. It doesn’t even need to be in the graph… and they would not be that close, especially between heating and cooling, I bet mine would be around 15F off or more.
I agree with you that this doesn’t have to become a nightmare involving data splitting. Just displaying the average indoor temp for each dataset would greatly enhance the analysis graph’s context.
The thermal delta to the outside is important since the resistance curve and “x °/hr” slope are completely temperature dependant. Even if the curve math is the same, the mechanics of heating at 62F and cooling at 77F are very different. It seems a little haphazard to act as though they are interchangeable because the range is “narrow,” especially when it comes to different seasons.
is not need to be a new series. For such run, a brief label like as “avg indoor: 64.3F” in the sidebar, tooltip, or legend would suffice. prevents fragmenting points and preserves the model while providing the user with a sanity check on where that slope genuinely resides on the curve.
I’ve also observed 10–15F differences between heat/cool averages, so yes, context is more important than most people realize, particularly when comparing runs that are months apart.
