配Seaborn

Question

I have a pandas dataframe that is composed like this:

User_id	Calls	Index
7A	8	19-05-2020
10B	5	19-05-2020
7A	2	20-05-2020
10B	6	20-05-2020

With in index the dates and times at which they make their calls even if here I put it on the right under pandas say that the dates are the index on the left.

I would like to make the distribution of the users of their number of calls that they make in the year.

Currently there are more than 400 different users so I will not do all of them only those who have the largest number of calls that I find easily with this: ret.groupby(['user_id'])['calls'].sum().sort_values(ascending=False).head(10)

I try to make a loop that takes the first 10 users who have the most calls and makes their distribution of calls in the year (so we have on the x-axis the months of the year written in number or not, on the y-axis the density or the number of calls, and we would have a title to say of which users we made the graph)

How can I do it?

Answer 1

try something like this, first you sort by your column 'calls' and after that you keep the first 10 rows

ret = ret.groupby(['user_id'])['calls'].sum().sort_values(ascending=False, by=['calls'])
ret = ret.iloc[:10]

Answer 2

IIUC try resample sum to get total calls per month then use plot :

*Note the provided data contains only a single month, so I added a second month to make the graph more clear:

import pandas as pd
from matplotlib import pyplot as plt

df = pd.DataFrame({'User_id': ['7A', '10B', '7A', '10B'], 'Calls': [8, 5, 2, 6],
                   'Index': ['19-05-2020', '19-05-2020', '20-06-2020',
                             '20-06-2020']}).set_index('Index')

df.index = pd.to_datetime(df.index, format='%d-%m-%Y')

users = ','.join(df['User_id'].unique())

plot_df = df.resample('1M')['Calls'].sum()
ax = plot_df.plot(kind='bar',
                  rot=0,
                  title=users,
                  ylabel='Calls',
                  xlabel='Months')

ax.set_xticklabels(plot_df.index.strftime("%Y-%m"))
plt.show()

---

df :

           User_id  Calls
Index                    
2020-05-19      7A      8
2020-05-19     10B      5
2020-06-20      7A      2
2020-06-20     10B      6

users :

7A,10B

plot_df :

Index
2020-05-31    13
2020-06-30     8
Freq: M, Name: Calls, dtype: int64