Not so easy with the enphase app in Homey. You need the cumulative kWh values for import and export.
If you can only use Watt you need to create seperate virtual devices first using GROUP app or advanced capabilities app. It has been mentioned multiple times in this forum so you can do a search on that.
Very nice this prediction feature 
When I enable this feature, are these values also added to the next hours JSON?
Based on which rules do you mark a bar high (orange) or low (green)? I suppose something like x % above average and x % under average? Could this x% be customisable?
Yes (but untested)
Simply the highest 4 and lowest 4 of all shown hours.
Tomorrow JSON also working
Just reporting back: this flow card works fine for me now.
Yes. But they are predictions, so just see what happens 
Maybe a lot of sun and/or wind is predicted in Germany tomorrow?
Maybe, but I very much doubt the big gap between 23:00 (actual price) and 00:00 (prediction).
And itâs Austria, not Germany
Sorry 
But looking at the graph from the source data, it is indeed showing a big gap. Would be fun if you could post the real prices graph later this afternoon to see the difference with the predicted prices.
Edit: this image is showing how well the predictions in Austria were for the last few days. Pretty well imo, just today your real prices are much higher than predicted. Someone is messing with your market prices
One bug found - although the App got the actual prices for tomorrow, it didnât refresh the chart for âNext hoursâ; it still shows the prediction. Refresh doesnât help.
But it shows the correct ones in âTomorrowâ:
I also saved the estimated prices as JSON - here together with the actual:
Estimated: {"0":0.1297,"1":0.1107,"2":0.1056,"3":0.1023,"4":0.1037,"5":0.1113,"6":0.1235,"7":0.1359,"8":0.1434,"9":0.1439,"10":0.1384,"11":0.1311,"12":0.1261,"13":0.1262,"14":0.1314,"15":0.1397,"16":0.1481,"17":0.1537,"18":0.1543,"19":0.1497,"20":0.1417,"21":0.1329,"22":0.1259,"23":0.1212}
Actual: {"0":0.1789,"1":0.1712,"2":0.1688,"3":0.168,"4":0.1689,"5":0.1715,"6":0.1685,"7":0.1735,"8":0.1756,"9":0.1711,"10":0.1682,"11":0.1651,"12":0.1607,"13":0.1575,"14":0.1682,"15":0.1756,"16":0.1917,"17":0.198,"18":0.1985,"19":0.1954,"20":0.1927,"21":0.1796,"22":0.1766,"23":0.1656}
Also for comparison here the estimates from AleaSoft. The line âWâ are the acutal prices, âW-1â is the estimate from one week ago.
It doesnât seem to be any material price distortion. Iâll keep an eye on it in the next days.
It shows the new prediction for Monday. So no bug
I just saw the big gap (again) and though it didnât change⌠should have looked more closely.
Which site is this graph (and your data) coming from?
I dont dare say. Iâm affraid access will be blocked if too many people start checking it.
Today I had estimated prices for tomorrow until around noon. Now (shortly after 13:00) those prices are gone and I only see prices for today. The option âEnable forecast pricesâ is enabled in the option. Any idea why?
I had the same, after 14u the prices for tomorrow and prediction for day after tomorrow were available.
So it looks like if prices are late, the prices are added at 14:0# and between 13:0# and 14:0# the prices for tomorrow and prediction is empty.
So it looks like if prices are late, the prices are added at 14:0# and between 13:0# and 14:0# the prices for tomorrow and prediction is empty.
My thought too. Seems to me that the estimations are removed right before it tries to get the actual prices for tomorrow; if it doesnât get them immediately, estimations remain empty until it does.
Just to keep a close eye on the gap between estimated and actual prices, I added a flow that compares them every day.
For todayâs prices I got yesterday:
â estimated: 0.1304 â actual: 0.1754 â %: 34.47 min diff: 0.0272 max diff: 0.0657
And for tomorrowâs prices I got today:
â estimated: 0.1559 â actual: 0.1966 â %: 26.07 min diff: 0.0258 max diff: 0.0672
The gap is pretty big. Lets see how this develops in the next days.
Do you know how old those estimations are?
Yeah need to look into that. Will be a couple of weeks probably beforeI have time to dig into it.
