CODEWITHSUNDEEP
pandasseries
user1315789
user1315789

Reputation: 3659

Get percentage of positive numbers in this panda series

I have a panda series that looks like this;

0     -10.93
1      -9.18
2      -6.87
3      7.90
4      9.18
5      6.87
6      9.12
7      9.12
8      8.75
9      7.75

I would like to get the percentage of positive numbers in this series. In this case, the answer should be 70%.

I am using python 3.6

Upvotes: 1

Views: 1217

Answers (2)

jezrael
jezrael

Reputation: 862911

Use mean of boolean mask:

a = s.gt(0).mean() * 100

a = (np.sign(s) == 1).mean() * 100

print (a)
70.0

If want also all negative, positive and zero values is possible use numpy.sign with value_counts:

print (s)
0   -10.93
1    -9.18
2    -6.87
3     7.90
4     9.18
5     6.87
6     9.12
7     9.12
8     8.75
9     0.00 <- changed last value to 0
Name: a, dtype: float64

b = np.sign(s).value_counts(normalize=True).rename({0:'zero', 1:'pos', -1:'neg'})
print (b)
pos     0.6
neg     0.3
zero    0.1
Name: a, dtype: float64

Upvotes: 2

Zero
Zero

Reputation: 76927

Use

In [381]: s.ge(0).sum() / len(s.index) * 100.
Out[381]: 70.0

Or

In [383]: 100. * (s >= 0).sum() / len(s.index)
Out[383]: 70.0

Upvotes: 1

Related Questions