Color and size per glyph

[slinnarsson]

This is great!

Feature request for scatter plots: would be fantastic if you could set the color and size of each glyph individually, for visualization of multidimensional scientific data.

[jeammimi]

I agree and started to implement it.

[maartenbreddels]

Thank you, as @jeammimi is saying, he's working on it in #10

[maartenbreddels]

Fixed by #10 thanks to @jeammimi

[deeptavker]

Hey, I'm trying to plot a 3D scatter with colorscale. Something like this :

I am using ipyvolume.pylab.scatter but I don;t understand how I can have colors in such fashion, I looked at #10 and tried it using the color option.

A trivial code which follows the idea that I'm using to achieve the above shown image.

import ipyvolume.pylab as p3
import numpy as np

#points
x = np.arange(10)
y = np.arange(10)
z = np.arange(10)

#colors
#For each point x,y,z the color corresponding to the same index in c should be used
c = np.arange(10).astype('float64')/10

p3.clear()
p3.scatter(x, y, z, color=np.meshgrid(range(3),c)[1], size=3)
p3.show()

The output :

Can I directly use c without using np.meshgrid() or something similar? Note that I used np.meshgrid just to create an array with required shape. I'm sure there is a cleaner way to do this.
How do I set a colorscale as shown in the first image? And how to add a colorbar legend?

Any help is appreciated, thanks :-)

[maartenbreddels]

Hi,

maybe #12 would have been useful, but otherwise the code below might help. The idea is to use a matplotlib colormap, that maps to rgb(a) values.

import ipyvolume.pylab as p3
import numpy as np
import matplotlib.cm
#points
x = np.arange(10)
y = np.arange(10)
z = np.arange(10)

#colors
colormap = matplotlib.cm.viridis
c = colormap(np.linspace(0, 1, len(x)))
p3.clear()
p3.scatter(x, y, z, color=c, size=3)
p3.show()

Hope this helps, let me know!

cheers,

Maarten

[deeptavker]

Hey! This works just fine :)

Yes, I should have looked at #12

Since I probably cannot use pyplot.colorbar for this, I'm still trying to find a way to create legends for multiple scatters on the same cell output display. Anyway, the matplotlib colormap technique works well for the colors.

Thanks a lot!

Although I'm new to this, it's quite simple to use and integrates really well with ipywidgets as compared to matplotlib. Awesome stuff :-)

[maartenbreddels]

Hi,

thanks!
Indeed, there isn't any colorbar support yet in ipyvolume. You can try to plot that in the same cell using matplotlib, make sure you call pylab.show(), although I've never plotted just a single colorbar it should be possible.

cheers,

Maarten