scatterplot axis autoscale fails for small data values

I'm finding that the scatterplot autoscaling fails when plotting small values, e.g. <0.001.

The autoscale axis range will never go below -0.008 to 0.008 for me, regardless of how small the range of the data being plotted.

plt.figure()
x = np.array([0,1,2,3])
x = np.array([2,4,5,9])
plt.scatter(x,y)

Data in a normal range produces

plt.figure()
x = np.array([0,1,2,3])
x = np.array([2,4,5,9])
plt.scatter(x/10000,y/10000)

While scaling it down produces

On 2.x it seems to go to pm 0.002.

I assume this is due to the non-singular expanding logic?

Fix matplotlib#6015: Support fallback font correctly

@mdboom

ping @mdboom, I may be wrong, but the referenced commit above doesn't relate to this bug?

Disable autoscalling in swarmplot

This problem is still present in current master.

Apart from manually scaling, is there a workaround for this issue?

You can use .plot() instead and use linestyle='none' and define markers. Obviously this doesn't have the same features as scatter but for basic functionality it works. Am Sa., 5. Jan. 2019, 01:52 hat Fraser Parlane <notifications@github.com> geschrieben:

…

Apart from manually scaling, is there a workaround for this issue? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#6015 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATppFUdtwF_UIINm1lIu8iEdDPDU0UV9ks5u__c1gaJpZM4HcOg4> .

If you still want the scatter functionality and have autoscale working, there's a good workaround I found on stackoverflow where you create an invisible plot after scatter:

import matplotlib.pyplot as plt

y=  [1.09E-05,  1.63E-05,   2.45E-05,   3.59E-05,   5.09E-05,   6.93E-05,   9.07E-05]
x=  [0, 10, 20, 30, 40, 50, 60]

fig3, ax3 = plt.subplots()
ax3.scatter(x, y, color='k', marker = "o")
ax3.plot(x, y, color='none')
ax3.relim()
ax3.autoscale_view()
plt.show()

Closing as fixed by #13642

Has this problem with 3.1.2.

#13642 will be in 3.2

I installed the 3.2.0rc3 package, and I can confirm that this fixes almost all the issues with autoscaling and scatter. However, if you have points with 0 for the y-axis and enable log scaling there is still an issue with the scatter plot auto scaling (that doesn't show up with the standard plot function).

import itertools

import numpy as np
from matplotlib import pyplot as plt


x_vals = [4.38462e-06,
          5.54929e-06,
          7.02332e-06,
          8.88889e-06,
          1.125e-05,
          1.42383e-05,
          1.80203e-05,
          2.2807e-05,
          2.88651e-05,
          3.65324e-05,
          4.62363e-05,
          5.85178e-05,
          7.40616e-05,
          9.37342e-05,
          0.000118632,
          ]

y_vals = [0.0,
          0.10000000000000002,
          0.182,
          0.332,
          0.604,
          1.1,
          2.0,
          3.64,
          6.64,
          12.100000000000001,
          22.0,
          39.60000000000001,
          71.3,
          ]

pts = np.array(list(itertools.product(x_vals, y_vals)))

fig = plt.figure("Scatter plot")
ax = fig.gca()
ax.set_xscale('log')
ax.set_yscale('log')
ax.scatter(pts[:,0], pts[:,1]) # This only shows four rows of points

fig = plt.figure("Regular plot")
ax = fig.gca()
ax.set_xscale('log')
ax.set_yscale('log')
ax.plot(pts[:,0], pts[:,1], marker="o", ls="") # This works

plt.show()

Can you open this as a new issue?

Navigation Menu

Repository navigation

scatterplot axis autoscale fails for small data values #6015

1 remaining item

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

scatterplot axis autoscale fails for small data values #6015

Description

Activity

tacaswell commented on Feb 18, 2016

dstansby commented on Jan 29, 2017

1 remaining item

ImportanceOfBeingErnest commented on Jul 12, 2018

FraserParlane commented on Jan 5, 2019

LindyBalboa commented on Jan 5, 2019

trbedwards commented on Jan 11, 2019

ImportanceOfBeingErnest commented on Dec 17, 2019

geyang commented on Jan 19, 2020

jklymak commented on Jan 19, 2020

moloney commented on Feb 14, 2020

jklymak commented on Feb 15, 2020

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions