Use own colormap in a plot according to a 3rd value (no scatterplot)

Question

I created my own colormap for matplotlib and I can use it for a scatterplot (thanks to stack overflow).
How can I use it in the normal plot function (plt.plot()). It does not accept the "cmap" parameter.
I want to plot x,y values and color them according to a z-value using my colormap and have a different marker according to the z-value.

can't be done, Line2D objects only have one color. Why won't scatter work for you? Can you show some code of what you are trying to do? — tacaswell
I agree with @tcaswell - show some code. If you have no ideas at all, I'd suggest trying a 2-step process, first, define markers and their color, then plot them individually. — Schorsch
@Schorsch I think that is what I had in mind. Create a list containing the colors and one the markers and then plot the points one after the other. But I am not sure how to convert the Z-value into a color from the colormap and use it. — Trollbrot

sissi_luaty sissi_luaty · Accepted Answer · 2013-06-03T20:28:20

There is no buit-in function to do what you want, so I've written a python script to answer your question (the comments explain the code):

import matplotlib.pyplot as plt
import numpy as np
from matplotlib import cm


fig = plt.figure()

width=4.467    # width in inches of my figure
height=4.344   # height in inches of my figure
DPI=90         # dots per inch

xmin=0
xmax=2*np.pi
xx_step=1.0*(xmax-xmin)/(width*DPI)
print "xx_step:", xx_step
xx=np.arange(xmin, xmax+xx_step, xx_step)

ymin=-1
ymax=1
yy_step=1.0*(ymax-ymin)/(height*DPI)
print "yy_step:", yy_step
yy=np.arange(ymin, ymax+yy_step, yy_step)

x_surf, y_surf = np.meshgrid(xx, yy)                   # generate a mesh
z_surf = np.zeros(x_surf.shape)+float('+inf')          # to paint with white color the regions that will not contain the function's curve (the default value depends on your colormap)

y_dots=np.sin(xx)
colors=[i for i in range(len(xx))]                     # colors of each point of the function, in this example they vary linearly

def find(my_list, func):
    # this find() method is from:
    # http://stackoverflow.com/questions/5957470/matlab-style-find-function-in-python
    return [j for (j, value) in enumerate(my_list) if func(value)]

def insert_dot(x, y, color):
    global z_surf
    def insert_except_at_contours(x, y, color):
        try:
            z_surf[y, x]=color
        except:
            pass
    insert_except_at_contours(x-1, y, color)
    insert_except_at_contours(x, y-1, color)
    insert_except_at_contours(x, y, color)
    insert_except_at_contours(x, y+1, color)
    insert_except_at_contours(x+1, y, color)

flag=0
delta_y=0
for i in range(len(xx)):   # for each pair of values in xx and the corresponding function that is stored in y_dots, find the corresponding indices in the mesh
    xx_index=i
    cur_array=abs(y_dots[i]-yy)
    cmp_test=min(cur_array)
    yy_index=find(cur_array, lambda x: (x==cmp_test))[0]
    if not flag:
        insert_dot(xx_index, yy_index, colors[i])
        flag=1
    else:       # this "else" part (and the corresponding flag) paints a straight line between the given dots, to improve the resolution (optional).
        delta_y=np.sign(last_yy_index-yy_index)
        if delta_y==0:
            delta_y=1
            last_yy_index+=1
        for j in range(yy_index, last_yy_index, delta_y):
            insert_dot(xx_index, j, colors[i])
    last_yy_index=yy_index

print "in pcolor()..."
obj=plt.pcolor(x_surf, y_surf, z_surf, cmap=cm.hot)     # plot the graph
print "done!"
plt.clim(min(colors), max(colors))
fig.colorbar(obj)

axis=plt.gca()
axis.set_xlim([min(xx), max(xx)])
axis.set_ylim([min(yy), max(yy)])

plt.show()

result:

Use own colormap in a plot according to a 3rd value (no scatterplot)

1 Answers