[Solved] Spiking in wind direction chart
Since a couple of weeks (since 3.7?) my wind dir charts look spikey. With a more or les regular interval the wind direction (from) spikes to North, 0°, then immediately back to the correct wind direction.
First i suspected it was the Davis hardware. The anemometer kit was even replaced under warranty... But that was not the solution.
Toninght I checked the shortcode, and found something strange; When i choose "Direction (From)" the spikes are visible, but when i choose "Direction (To)" they magically disappear... See the 2 attached screenshots.
It's absolutely normal. And I know it looks ugly but the values are good. It's because direction is source +/- 180° (that's the opposite angle), and when you show circular values (like angles) in a linear representation you obtain that if you have values which are near 0, 180 and 360° (it's like quick oscillations).
To confirm your values are correct, just generate a windrose with source and another with direction you will see it's the same windrose "mirrored", it's because, this time, the representation is circular...
In fact, lines and multilines are not help to visualize circular data 😉
Thanks for your quick response!
But somehow these spikes where not present before. Those datapoints are always exactly 0°. So it is not the wind vane passing north. The North datapoints are not measured datapoints. I'll somehow have to check if Meteobridge reports that data. 🤔
This graph is very useful for displaying change in wind direction during the day.
Yes, it's because some stations are measuring direction, other are measuring source. Just make the test: forget lines and display only dots, you will see what I'm speaking of...
Don't be sorry, that's the most painful thing in data visualization (and the most difficult thing to "read"): circular values in a linear visualization... 🙂
Well, it turns out the anemometer kit was indeed defective last week, and did need replacement for losing contact. And losing contact means it reports 0°. Last's behaviour looked the same, even with wind from the south. That caused spiking, caused by the defective anemometer. I checked the Meteobridge files and now i see the difference. 2 different things causing a similar effect😳
This time spiking is "the real thing" and no error. Followed your advice, and changed the line into dots👍