diff --git a/source/core/numpy_scipy.rst b/source/core/numpy_scipy.rst
index 8f35455..95ad5d2 100644
--- a/source/core/numpy_scipy.rst
+++ b/source/core/numpy_scipy.rst
@@ -458,7 +458,7 @@ image.
 
   import scipy.signal
   img_sm = scipy.signal.medfilt(img, 5)
-  sigma = median(err)
+  sigma = np.median(err)
   bad = np.abs(img - img_sm) / sigma > 8.0
   img_cr = img.copy()
   img_cr[bad] = img_sm[bad]
diff --git a/source/installation/python_install.rst b/source/installation/python_install.rst
index 34d3acd..9b43e45 100644
--- a/source/installation/python_install.rst
+++ b/source/installation/python_install.rst
@@ -19,11 +19,12 @@ You may need to choose between 32-bit and 64-bit installations.  Generally
 speaking you should choose 64-bit, but read `64 versus 32 bit`_ for some caveats
 or if you aren't sure if your CPU is 64-bit.
 
-For this workshop you can use either Python 2.6, 2.7 or Python 3 (version >=
-3.3).  On the general question of whether to use Python 2 or Python 3, at this
-point the major package support for both is quite similar, and (as of
-early 2015) it appears that in the overall community Python 3 usage is
-becoming substantial.
+For most of this workshop you can use either Python 2.7 or Python 3 (version >=
+3.3).  The exception is the Sherpa fitting tutorial which can only be done with
+Python 2.7.  On the general question of whether to use Python 2 or Python 3, at
+this point the major package support for both is quite similar, and (as of
+early 2015) it appears that in the overall community Python 3 usage is becoming
+substantial.
 
 .. _`anaconda_option`:
 
@@ -113,6 +114,7 @@ core scientific Python installation, do the following and check version numbers:
   import numpy
   import scipy
   import scipy.linalg
+  import matplotlib
   import matplotlib.pyplot as plt
 
   print(numpy.__version__)
@@ -120,7 +122,7 @@ core scientific Python installation, do the following and check version numbers:
   print(matplotlib.__version__)
 
   x = numpy.linspace(0, 20, 100)
-  plt.plot(x, sin(x))
+  plt.plot(x, numpy.sin(x))
   print(scipy.linalg.eig([[1,2],[3,4]]))
 
 The commands above should succeed with no errors.  The version numbers should
diff --git a/source/intro/quick-tour.rst b/source/intro/quick-tour.rst
index f1a0480..dd48bf0 100644
--- a/source/intro/quick-tour.rst
+++ b/source/intro/quick-tour.rst
@@ -182,7 +182,7 @@ including convolution with a Gaussian filter and the addition of noise.
   image += random.normal(3., 0.01, image.shape)
 
   # Write out to FITS image
-  pits.writeto('cluster.fits', image, clobber=True)
+  fits.writeto('cluster.fits', image, clobber=True)
 
 The simulated cluster image is below:
 
diff --git a/source/plotting/matplotlib.rst b/source/plotting/matplotlib.rst
index 5b04eee..69c458b 100644
--- a/source/plotting/matplotlib.rst
+++ b/source/plotting/matplotlib.rst
@@ -704,7 +704,7 @@ these arguments are ``(x,y)`` tuples::
 
   plt.clf()
   t = np.arange(0.0, 5.0, 0.01)
-  s = np.cos(2*pi*t)
+  s = np.cos(2*np.pi*t)
   lines = plt.plot(t, s, lw=2)
 
   plt.annotate('local max', xy=(2, 1), xytext=(3, 1.5),