Use 10 minute average instead of instantaneous data

[skalnik]

Someone requested we use the 10 minute average instead of the latest value. I think that'd be good, but I believe we'll need to poll ThingSpeak instead of PA for that. Unsure how hard that'd be to implement, but I wanted to make note of it.

[obra]

Is there a -reason- they want the 10 minute averages?

One -possible- alternative to using the thingspeak API would be to compute our own rolling 5 or 10 minute average if we have the data and the sensor we're polling hasn't changed.

[obra]

(I don't believe that there's no reason to want 10 minute averages or that those averages aren't better. I am absolutely not skilled in that side of things. But I'm a big fan of documenting the "why" of changes.

[skalnik]

They said they preferred the 10 minute average to avoid sudden spikes/dips and smooth those out. I think that's fair, but also not a huge problem. I promised I'd document this but I'm not sure it's a huge priority.

[obra]

If we end up going down that path, it might be interesting to also look at averaging data from multiple sensors, ideally ones we're in the middle of. But that may be a really bad idea in non-densely-covered areas.

…

On Thu, Sep 10, 2020, at 2:44 PM, Mike Skalnik wrote: They said they preferred the 10 minute average to avoid sudden spikes/dips and smooth those out. I think that's fair, but also not a huge problem. I promised I'd document this but I'm not sure it's a huge priority. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#50 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALC2AGKJWEALYYUA67JU3SFFCFPANCNFSM4REMNFEQ>.