CODEWITHSUNDEEP
pythonmatplotlibplotcolorsinterpolation
Euler_Salter
Euler_Salter

Reputation: 3571

plot two interpolations of two set of points in the same plot

I have a set of points which can be "distinguished" in two parts, based on their coordinates. To make an example, imagine:

np.random.seed(1234)           #I hope this is how you use random seed, not sure

### First part of data ###
xcoord1 = np.random.uniform(10,20,100).reshape(-1,1)
ycoord1 = np.random.uniform(20,30, 100).reshape(-1,1)
part1 = np.hstack((xcoord1, ycoord1))
### second part of data ###
xcoord2 = np.random.uniform(60,90,100).reshape(-1,1)
ycoord2 = np.ranodom.uniform(5,25,100).reshape(-1,1)
part2 = np.vstack((xcoord2, ycoord2))

Important: the two part of points are totally disjoint as you can see here image

What I want to do

I have a set of values 

values1 = np.random.uniform(0,1, size = 200) #for part1 
values2 = np.random.uniform(0,1, size = 200) #for part2

each corresponding to the points in part1 and part2. I can easily plot them based on the color (here)

colors1 = [cmap(tl) for tl in values1]
colors2 = [cmap(tl) for tl in values2]
colors = ['yellow', 'red', 'green']
cmap = matplotlib.colors.ListedColormap(colors)
plt.scatter(part1[:,0], part1[:,1], c = colors1)
plt.scatter(part2[:,0], part2[:,1], c=colors2)

obtaining the following image1

Now I would like to do some interpolation on the colors, BUT I would like to have the color meshing only in the two areas where I have the points, not in between them! So I want to do two interpolations.

I managed to do the following:

xi = np.linspace(min(min(part1[:,0]), min(part2[:,0])), max(max(part1[:,0]), max(part2[:,0])), 200)
yi = np.linspace(min(min(part1[:,1]), min(part2[:,1])), max(max(part1[:,1]), max(part2[:,1])), 200)
xx, yy = np.meshgrid(xi, yi)
gridPoints = (xx,yy)
totalpart = np.vstack((part1, part2))
totalvalue = np.hstack((values1, values2))
zi = griddata(totalpart, totalvalue, gridPoints, method='linear')
plt.imshow(zi, origin = 'lower', cmap = cmap)

Which gives me

image3

Now this is an interpolation, but as you see the colors occupy the part of the plot in between the two "clusters", that used to be white. I would like to have each one of the two "clusters" (i.e. part1 and part2 ) to have their own interpolation (with the same colors if possible).

How can I do that?

WHAT I WANT, DONE IN PAINT

I tried to do in paint what I want to achieve, so you have a better idea:

imagepaint

Upvotes: 0

Views: 77

Answers (1)

Francesc Torradeflot
Francesc Torradeflot

Reputation: 126

I would create a mask to set the values of zi outside the two regions to np.nan. That is

mask1 = (gridPoints[0] >= 10) & (gridPoints[0] <= 20) & \
    (gridPoints[1] >= 20) & (gridPoints[1] <= 30)
mask2 = (gridPoints[0] >= 60) & (gridPoints[0] <= 90) & \
    (gridPoints[1] >= 5) & (gridPoints[1] <= 25)
mask = mask1 | mask2

zi[~mask] = np.nan
plt.imshow(zi, origin = 'lower', cmap = cmap)

The result looks like this: enter image description here

Upvotes: 1

Related Questions