Hi @netotz,

First you have to define a 2d quiver plot on map.

Here is a simple example of 2d quiver plot:

```
import plotly.figure_factory as ff
from numpy import sin, cos, pi
import numpy as np
x, y = np.mgrid[0:1:16j,0:1:16j]
z = 0.0*np.ones(x.shape)
u = sin(x*pi) * cos(y*pi) * cos(z*pi)
v = cos(x*pi) * sin(y*pi) * cos(z*pi)
fig = FF.create_quiver(x, y, u, v,
scale=0.065,
arrow_scale=.3,
angle=pi/18,
name='quiver',
line=dict(width=1, color='#8f180b'))
fig.show()
```

Inspecting `fig.data[0]`

you’ll notice that the corresponding trace is a Scatter trace, where x, y contain the end point coordinates for each segment that defines a quiver, separated by `None`

.

At this link you may download a zip archive that contains:

- the json file for the 2d figure for the flat Earth with countries and coastlines, a heatmap that has as z-values mapped to the colorscale the wind speed at each point, as well as the quivers defined as in the example above, but with x, y having as coordinates the geographic coords, lon and lat.
- a jupyter notebook where data from the 2d figure are mapped onto the sphere to get the 3d version for wind (the 3d plot you posted here).

The jupyter notebook for the 2d plot is no more available, that’s why I included in the archive only that for the 3d version.

Hope that this information can help.