I am a newbie to Data science and Machine Learning.

I am trying to understand the Normalization of the Dataframe values. Here is the scenario from the famous disaster i.e. Titanic and here is the code and result from a query:

dftitanic.groupby('Fsize')['Survived'].value_counts(normalize=False).reset_index(name='perc')

Result:

    Fsize  Survived  perc
0       1         0   374
1       1         1   163
2       2         1    89
3       2         0    72
4       3         1    59
5       3         0    43
6       4         1    21
7       4         0     8
8       5         0    12
9       5         1     3
10      6         0    19
11      6         1     3
12      7         0     8
13      7         1     4
14      8         0     6
15     11         0     7

And if I use .value_counts(normalize=True), the result would be:

dftitanic.groupby('Fsize')['Survived'].value_counts(normalize=True).reset_index(name='perc')

    Fsize  Survived      perc
0       1         0  0.696462
1       1         1  0.303538
2       2         1  0.552795
3       2         0  0.447205
4       3         1  0.578431
5       3         0  0.421569
6       4         1  0.724138
7       4         0  0.275862
8       5         0  0.800000
9       5         1  0.200000
10      6         0  0.863636
11      6         1  0.136364
12      7         0  0.666667
13      7         1  0.333333
14      8         0  1.000000
15     11         0  1.000000

And the data from describe():

        Fsize   Survived    Perc
count   16.0000 16.000000   16.000000
mean    4.6875  0.437500    55.687500
std     2.7500  0.512348    95.378347
min     1.0000  0.000000    3.000000
25%     2.7500  0.000000    6.750000
50%     4.5000  0.000000    15.500000
75%     6.2500  1.000000    62.250000
max     11.0000 1.000000    374.000000

My effort:

From /a/41532180, I got the following methods:

1. normalized_df=(df-df.mean())/df.std()

2. normalized_df=(df-df.min())/(df.max()-df.min()

However, from the results of describe(), the above two methods not matching the results of .values_counts(normalize=True).

A similar formula and description is present here: but did give understandable results.

Question:

How this Normalization being done? i.e. .value_counts(normalize=True)

Thanks a bunch.