Set precision on custom ticks with scientific style

Question

I have produced this code to generate the following graph:

import matplotlib.pyplot as plt
import numpy as np
x = np.linspace(1/10000, 1/2000, 13)
y = x**2
plt.plot(x, y, 'ro')
plt.ticklabel_format(style='sci', axis='x', scilimits=(0,0), useMathText=True)

I would like to set the xticks at the position of the data. If I then do plt.xticks(x, rotation=45) I get the ticks at the desired locations but with too many decimal places (see next picture). How do I get the ticks at the specified locations but with a controllable precision?

score 1 · Accepted Answer · answered Feb 25 '19 at 16:10

In order to get a predefined format for the ticklabels while maintaining the scientific multiplier you can use a simplified version of the OOMFormatter from my answer here.

import matplotlib.pyplot as plt
import numpy as np
import matplotlib.ticker

class FFormatter(matplotlib.ticker.ScalarFormatter):
    def __init__(self, fformat="%1.1f", offset=True, mathText=True):
        self.fformat = fformat
        matplotlib.ticker.ScalarFormatter.__init__(self,useOffset=offset,useMathText=mathText)
    def _set_format(self, vmin, vmax):
        self.format = self.fformat
        if self._useMathText:
            self.format = '$%s$' % matplotlib.ticker._mathdefault(self.format)

x = np.linspace(1/10000, 1/2000, 13)
y = x**2
plt.plot(x, y, 'ro')
plt.xticks(x, rotation=45)

fmt = plt.gca().xaxis.set_major_formatter(FFormatter(fformat="%1.1f"))
plt.ticklabel_format(style='sci', axis='x', scilimits=(0,0), useMathText=True)

plt.show()

score 0 · Answer 2 · answered Feb 25 '19 at 15:31

I managed to solve the issue by manually specifying the scientific limits and then creating custom labels:

import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(1/10000, 1/2000, 13)
y = x**2
e = 4

plt.plot(x, y, 'ro')
# plt.ticklabel_format(style='sci', axis='x', scilimits=(0,0), useMathText=True)
plt.xticks(x, ['{:.1f}'.format(10**e*s) for s in x])
plt.text(1.01, 0, 'x$10^{{:d}}$'.format(e), transform=plt.gca().transAxes)

This solves the issue though it requires manually specifying the exponent.

score 0 · Answer 3 · answered Feb 25 '19 at 16:08

A simple way to do this is to round the x values in the ticks assignment:

plt.xticks(np.round(x,decimals=6))

works for me for example.

Note that this will actually move the ticks to the indicated positions. With only one decimal this will be clearly visible.

Set precision on custom ticks with scientific style

3 Answers3

Linked