Wind direction graphical representation

[benoit-jpi]

In the graphical representation, the wind direction should not be interpolated or smoothed. It would be more appropriate to represent the direction of the wind in a stationary way on the time segments

[claustres]

To go further in this direction I think we actually need a specific representation for wind direction. Now we can easily visualize wind direction change at probed location with the wind barb marker by changing the current time in the timeline. However, the timeseries graph is not really readable for wind direction. A better representation could be based on oriented arrows, maybe as a collapsible subwidget in the the timeseries widget.