是否有可能在Python中的Radar-Chart上反转轴方向？

Question

I am trying to plot a Radar-Chart using the following code from the this source .

And my goal is, to reverse the r-Axis without having to remap my data points as my data is on a scale from 1 to 5, with 1 indicating very food and 5 very bad. (so I would loose the meaning of the scale, when reversing the datapoints)

(Which has been described here )

My first approach was to use matplotlibs inherent functionality .

So with the source being

# Draw ylabels
ax.set_rlabel_position(0)
plt.yticks([10,20,30], ["10","20","30"], color="grey", size=7)
plt.ylim(0,40)

My approach would be

# Draw ylabels
ax.set_rlabel_position(0)
plt.yticks([30,20,10], ["30","20","10"], color="grey", size=7)  # Reversed labels
plt.ylim(40,0) # Reversed axis, as described above

But the problem is, that the lower code never finishes. So i don't even know how to debug it, as i don't get any errors.

I also can't seem to reverse only the Axis labels (as with that approach it would be doable to just reverse the data and the labels)

Answer 1

If you want to go with the reverse labels thing, you have to use plt.yticks([10,20,30], ["30", "20", "10"], ...) since the first parameter corresponds to axes' values, and as you have not yet inverted it, they should remain in that order.

I checked on the plt.ylim inversion and it does end for me, but throws a rather cryptic error posx and posy should be finite values . Taking into account posx and posy are not parameters for this functions, there must be an underlying function that does not like this. In addition, having tested this for a non-polar plot, I guess the problem comes from the polar coordinates.

Looking around, I found both a github issue and an SO question , which resulted in a PR and posterior merge in Dec 2018. It should be available and work if you have the latest matplotlib version.

Answer 2

Have a look at the following... hopefully there's something you can use here. The way I got this to work was plotting rmax-r instead of r . I also reversed the order of the ticks, but kept the tick labels the same.

# Set up the data for plotting.
N=20
angles = 2.0*pi*np.linspace(0,1,N)
rmin = 0
rmax = 10
radii = rmax*np.random.random(N)

# Plot the non-reversed plot
plt.figure()
ax = plt.subplot(111,polar = True)
ax.plot(angles,radii)
ax.fill(angles, radii, 'b', alpha=0.1)
n_labels = 5
ticks1 = np.linspace(rmin,rmax,n_labels)
labels = [str(t) for t in ticks1]
plt.yticks(ticks1, labels)
plt.ylim(rmin,rmax)

# Reverse the plot
r2 = radii.max()-radii
plt.figure()
ax = plt.subplot(111,polar = True)
ax.plot(angles, r2)
ticks2 = np.linspace(rmax,rmin,n_labels)
labels = [str(t) for t in ticks1]
plt.yticks(ticks2, labels)
ax.fill_between(angles,r2,rmax,color='b',alpha = 0.1)
plt.ylim(rmin,rmax)