No Numeric Type To Aggregate

Question

I have a dataframe in which there is the following columns:

Date - Seller - Amount

Code sample:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
testframe = pd.DataFrame({'date': ['01/02/2019', '01/02/2019', '01/02/2019', 
'02/02/2019','02/02/2019','03/02/2019'], 
'Seller': ['John', 'Ada', 'John', 'Ada', 'Ada', 'Ada'], 'Amount': 
[150,200,158,200,60,90]})

I aggregated the data so:

agg=pd.DataFrame(base.groupby(['date','Seller'])['Amount'].sum())
agg.reset_index(inplace=True)

Then I tried, using FacetGrid, to display the Amount along the days per Seller(each row from FacetGrid would be a row). Like so:

g=sns.FacetGrid(data=agg,row='Seller')
g.map(sns.lineplot, 'Amount', 'date')

But I get the following error:

No numeric types to aggregate

There was another post here on Stackoverflow, but didn't quite give a me clue about what to do in order to solve.

I checked the variables from packages and it returned numpy.int64. I tried to convert'em to float and int using .astype() but didn't work too.

Environment: Jupyter Notebook

Can someone shed a light on this?

Thanks in advance

Please provide a [mcve] when asking about undesired behaviour in code. — ImportanceOfBeingErnest, Jun 07 '19 at 16:25
Sorry ! Didn't know it was a good practice. I'll edit the question — Lucas Abreu, Jun 07 '19 at 16:31

score 1 · Accepted Answer · edited Aug 17 '19 at 20:21

Possibly you mean something like this:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt


base = pd.DataFrame({'date': ['01/02/2019', '01/02/2019', '01/02/2019', 
                                   '02/02/2019','02/02/2019','03/02/2019'], 
    'Seller': ['John', 'Ada', 'John', 'Ada', 'Ada', 'Ada'], 'Amount': 
            [150,200,158,200,60,90]})


base["date"] = pd.to_datetime(base["date"])

agg=pd.DataFrame(base.groupby(['date','Seller'])['Amount'].sum())
agg.reset_index(inplace=True)


g=sns.FacetGrid(data=agg, row='Seller')
g.map(sns.lineplot, 'date', 'Amount')

plt.show()

Note that John's plot is empty because he only sold something on a single day, and one cannot draw a line from a point to itself.

I forgot to use the pd.to_datetime! Thanks a lot for the help — Lucas Abreu, Jun 07 '19 at 17:21

No Numeric Type To Aggregate

1 Answers1