如何在pandas系列中找到groupby函数的比率

Question

I have used groupby to group the dataset by the occupations and gender. Now I want to find the ratio between male and females for each occupation. I am unable to think of how to proceed.

Answer 1

Here's one way using pandas.pivot_table and vectorised Pandas calculations. Note this method removes the need to perform a separate groupby .

df = pd.DataFrame([['A', 'F'], ['A', 'F'], ['A', 'M'], ['B', 'M'], ['B', 'M'], ['B', 'F'],
                   ['C', 'M'], ['C', 'M'], ['D', 'F']], columns=['Occupation', 'Gender'])

# pivot input dataframe
res = df.pivot_table(index='Occupation', columns='Gender', aggfunc='size', fill_value=0)

# calculate ratios
sums = res[['F', 'M']].sum(axis=1)
res['FemaleRatio'] = res['F'] / sums
res['MaleRatio'] = res['M'] / sums

print(res)

Gender      F  M  FemaleRatio  MaleRatio
Occupation                              
A           2  1     0.666667   0.333333
B           1  2     0.333333   0.666667
C           0  2     0.000000   1.000000
D           1  0     1.000000   0.000000

Answer 2

Maybe quite late to the party but here's what I believe is the exact answer:

# create pivot
male_ratio = users.pivot_table(index='occupation', columns='gender', aggfunc='size', fill_value=0)

# calculate male ratio
sums = male_ratio[['F', 'M']].sum(axis=1)
male_ratio['MaleRatio'] = round(100 * male_ratio['M'] / sums , 1)

# result
male_ratio['MaleRatio']

occupation
administrator     54.4
artist            53.6
doctor           100.0
educator          72.6
engineer          97.0
entertainment     88.9
executive         90.6
healthcare        31.2
homemaker         14.3
lawyer            83.3
librarian         43.1
marketing         61.5
none              55.6
other             65.7
programmer        90.9
retired           92.9
salesman          75.0
scientist         90.3
student           69.4
technician        96.3
writer            57.8
Name: MaleRatio, dtype: float64

Answer 3

x=users.groupby(['occupation','gender'])['gender'].count()
    y=users.groupby(['occupation'])['gender'].count()
    r=((x/y)*100).round(2)
    print(r)

#ratio rule "x" is a count of gender(male/female), "y" is the total count of gender

occupation     gender
administrator  F          45.57
               M          54.43
artist         F          46.43
               M          53.57
doctor         M         100.00
educator       F          27.37
               M          72.63
engineer       F           2.99
               M          97.01
entertainment  F          11.11
               M          88.89
executive      F           9.38
               M          90.62