Reformatting pandas table - do I want a pivot?

Question

I'm sure this is quite simple, but my brain is frozen and there are so many different pivot and transpose methods. A hint would be nice at this stage.

I have this dataframe:

I want this:

I know how to get to here, if that helped, but I'm not sure if it does

FYI - The actual data has more columns and I need to separate out these four based on the "site" column, reformat everything, calculate some percentages, put the pieces back together, and eventually end up with something like this:

I'm hoping that if I can get on the right track for reformatting part of the data, I can repeat the process...

(then I need to figure out how to run a Chi-square test, but that's for later... :-(

Answer 1

The easiest resolution is df.stack :

df = pd.DataFrame({'MIC-m': [138, 3, 22, 45],
                   'MIC-t': [34, 90, 30, 53],
                   'MIC-q': [73, 13, 53, 68],
                   'Total': [229, 229, 229, 229]}, index=['H', 'L', 'M', 'X'])

# Drop total, because we need sum of columns, not rows
df.drop(columns='Total', inplace=True)

# Get final result
df = pd.DataFrame(df.append(df.sum().rename('Total')).T.stack(), columns=['count'])

yields:

             count
MIC-m H        138
      L          3
      M         22
      X         45
      Total    208
MIC-t H         34
      L         90
      M         30
      X         53
      Total    207
MIC-q H         73
      L         13
      M         53
      X         68
      Total    207