fixes division by zero error for kurt()

[mortada]

Currently if kurt() is called on a Series with equal values, it throws a ZeroDivisionError

In [1]: import pandas as pd
In [2]: import numpy as np
In [3]: s = pd.Series(np.ones(5))
In [4]: s.kurt()
ZeroDivisionError: float division by zero

This is not consistent with the case when kurt() is called on a DataFrame of equal values

In [5]: df = pd.DataFrame(np.ones((5, 5)))
In [6]: df.kurt()
Out[6]:
0    0
1    0
2    0
3    0
4    0
dtype: float64

with this patch s.kurt() will return 0 instead of throwing.

[shoyer]

This looks good to me.

@mortada could you please add a note to the what's new for 0.16? You can reference the issue number for this PR.

[mortada]

@shoyer sure I've added a note to whatsnew for 0.16, this PR is updated.

[jreback]

is skew ok? are there similar tests? if not can you add

[mortada]

@jreback skew() doesn't have this problem.

In [4]: for i in range(1, 5):
    print 'i is %d, skew is %f' % (i, pd.Series(np.ones(i)).skew())
   ...:
i is 1, skew is nan
i is 2, skew is nan
i is 3, skew is 0.000000
i is 4, skew is 0.000000

and sure I will add similar tests to skew

[mortada]

@jreback PR is updated, please take a look

[shoyer]

Lgtm

[shoyer]

Thanks!