You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I would like to record multiple devices in the house with HexaBus plugs+ and plot a sum graph with a minute resolution. The fact that each plug submits its readings at different seconds is not quite good for this use case. My quick and dirty workaround was:
just cut off split seconds of timestamp
look at the 5 values before and after the current value to see if there's another reading for the same minute
average over values of the same minute
Of course, it would be even better to calculate a more accurate minute value by taking the last value before and the first after the minute in question.
E.g., if you had readings (UNIX timestamp here converted to human-readable datetime)
imagine a line through the points (2012-10-01 19:59:40;50) and (2012-10-01 20:00:30;100) which could also be referred to by (-20;50) and (30;100) when counting the seconds to the next full minute (full minute = 0)
the linear equation has the form y = ax + b
the gradient of the line through the two points is a = (100-50)/((30-(-20)) = 50/50 = 1
with the point (-20;50) you get b=70
do the same with the next two timestamps (-30;100) and (10;200), you get a = 2,5 and b = 175
That way, the normalized readings would read:
datetime value
2012-10-01 20:00:00 70
2012-10-01 20:01:00 175
Could we have this calculation in libklio?
Cheers,
Monika
The text was updated successfully, but these errors were encountered:
NOTE: This was originally submitted here: mysmartgrid/hexabus#107
Hi,
I would like to record multiple devices in the house with HexaBus plugs+ and plot a sum graph with a minute resolution. The fact that each plug submits its readings at different seconds is not quite good for this use case. My quick and dirty workaround was:
Of course, it would be even better to calculate a more accurate minute value by taking the last value before and the first after the minute in question.
E.g., if you had readings (UNIX timestamp here converted to human-readable datetime)
datetime value
2012-10-01 19:59:40 50
2012-10-01 20:00:30 100
2012-10-01 20:01:10 200
do the same with the next two timestamps (-30;100) and (10;200), you get a = 2,5 and b = 175
That way, the normalized readings would read:
datetime value
2012-10-01 20:00:00 70
2012-10-01 20:01:00 175
Could we have this calculation in libklio?
Cheers,
Monika
The text was updated successfully, but these errors were encountered: