It does reflect the oddness I saw when I tested having a different timezone in my browser (ie. the PC used for viewing the app) to the timezone defined in emoncms. It looks like the daily usage numbers in the first screenshot almost match the power graph in the second and vice-versa?
When you define the start times, you should always start at 0, so if you’re “Day” is between 10pm and midnight GMT and you wanted to shift everything by an hour, add 0:1 to the start of the updated definition:
0:1,1:0,6:1,14:0,17:1,21:0,23:1
I don’t think that will make much (if any) difference to the issue you see though
As TrystanLea pointed out, the API call used for the daily graphs uses the emoncms timezone in its calculations, whereas the power detail graph uses the browser timezone. I did start looking at doing the daily usage graph differently (essentially using the same API call being used by the detail graph), but my day job got in the way. That change should fix the daily use to at least match the power detail graph.
I might get a chance to look at it in the next day or so (I think I was almost done re-writing it), I’ll re-post here when I do.
The only difference is that for screenshots 3 & 4 I have advanced my tier start times by 1 hour, so screenshot 3 &4 should should show similar results to screenshots 1 &2. The power graphs display correctly; it is the bar chart (screenshot 3) that is awry - I would expect it to be showing an even higher reading for off-peak use (>80%) than can be seen in screenshot 1, but instead it is showing around 50%.
Sorry, I just realised I wasn’t clear in my response, I had assumed (incorrectly) 2 screenshots in your post when obviously there has to be 4!
Should have said:
It looks like the daily usage numbers in the first screenshot almost match the power graph in the 4th screenshot and the daily usage numbers from the 3rd screenshot match the power graph in the 2nd screenshot?
The 3rd and 2nd screenshots are closer, but looking at the second, I would still expect the off-peak use (light blue) to be significantly greater than the Day (dark blue). But the third screenshot is definitely paired with the fourth screenshot anyway - I have not muddled the screenshots up! The power graphs are displaying correctly, so one would expect the first screenshot to look similar to the fourth, although the off-peak use should be even less. The problem is that the bar graph data in screenshot 3 does not reflect the usage as displayed in the screenshot 4 power graph. In both sets the underlying usage data is the same: all that has changed is the times for the tiers. The data in the powergraphs is from the Aug 15 bar in both cases.
I am not sure who to speak to but I am also interested in having Sunday as an offpeak day. Here in South Africa our tier times changes in winter and summer but this winter (June -August) our times is as follows:
Tier names and costs:
OffPeak: 0.8003, Standard: 1.3976, Peak:3.6469