Error with matplotlib when running examples of scikit learn

Question

I am recently learning python, and I came across a package called scikit learn where we can use python libraries and custom made codes to generate various plots. I am already have installed all the dependancies and then I have downloaded and installed the scikit learn but when I am trying to run the example codes I am getting the error in generating the plot.

code

from sklearn import datasets
from sklearn.cross_validation import cross_val_predict
from sklearn import linear_model
import matplotlib.pyplot as plt

lr = linear_model.LinearRegression()
boston = datasets.load_boston()
y = boston.target

# cross_val_predict returns an array of the same size as `y` where each entry
# is a prediction obtained by cross validated:
predicted = cross_val_predict(lr, boston.data, y, cv=10)

fig,ax = plt.subplots() ## error comes from calling the function plt.subplots()
ax.scatter(y, predicted)
ax.plot([y.min(), y.max()], [y.min(), y.max()], 'k--', lw=4)
ax.set_xlabel('Measured')
ax.set_ylabel('Predicted')
fig.show()

Error

fig,ax = plt.subplots()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/pymodules/python2.7/matplotlib/pyplot.py", line 866, in subplots
    fig = figure(**fig_kw)
  File "/usr/lib/pymodules/python2.7/matplotlib/pyplot.py", line 343, in figure
    **kwargs)
  File "/usr/lib/pymodules/python2.7/matplotlib/backends/backend_tkagg.py", line 80, in new_figure_manager
    window = Tk.Tk()
  File "/usr/lib/python2.7/lib-tk/Tkinter.py", line 1712, in __init__
    self.tk = _tkinter.create(screenName, baseName, className, interactive, wantobjects, useTk, sync, use)
_tkinter.TclError: no display name and no $DISPLAY environment variable

Is this something to do with setting up the environmental variable? I am using python 2.7.3 and this package should work with it or doing I have to put the path or something? sorry am very new to python just quickly trying to figure out if I can use this package for generating plots quickly or not. Any help to get rid of this error is appreciated.

Answer 1

I suspect that you are trying to run your code on a distant machine ( ssh maybe) or simply matplotlib is not well installed.

If you are in the first case, I would suggest to use savefig with the command matplotlib.use('Agg') that tells matplotlib to not use the default display. This will produce a .png file that you could import it to your home using scp :

import matplotlib
matplotlib.use('Agg')
from pylab import savefig
from sklearn import datasets
from sklearn.cross_validation import cross_val_predict
from sklearn import linear_model
import matplotlib.pyplot as plt

lr = linear_model.LinearRegression()
boston = datasets.load_boston()
y = boston.target

# cross_val_predict returns an array of the same size as `y` where each entry
# is a prediction obtained by cross validated:
predicted = cross_val_predict(lr, boston.data, y, cv=10)

fig,ax = plt.subplots() ## error comes from calling the function plt.subplots()
ax.scatter(y, predicted)
ax.plot([y.min(), y.max()], [y.min(), y.max()], 'k--', lw=4)
ax.set_xlabel('Measured')
ax.set_ylabel('Predicted')

# don't use fig.show()

savefig('FIGURE_NAME.png') # savefig from matplotlib