I have a data frame with df
name :
>>> df.head()
InvoiceNumber | ProductCode | InvoiceDate | UnitPrice | CustomerId | Country | |
---|---|---|---|---|---|---|
0 | 489434 | 85048 | 2009-12-01 07:45:00 | 6.95 | 13085 | United Kingdom |
1 | 489434 | 79323P | 2009-12-01 07:45:00 | 6.75 | 13085 | United Kingdom |
2 | 489434 | 79323W | 2009-12-01 07:45:00 | 6.75 | 13085 | United Kingdom |
3 | 489434 | 22041 | 2009-12-01 07:45:00 | 2.1 | 13085 | United Kingdom |
4 | 489434 | 21232 | 2009-12-01 07:45:00 | 1.25 | 13085 | United Kingdom |
I want a bar plot of my data frame based on number of sold product in each day of week, like this :
I’m use this code :
fig1 ,ax1 = plt.subplots(figsize=(15, 6)) df["dayName"] = pd.to_datetime(df["InvoiceDate"]).dt.day_name() df.groupby("dayName").size().plot(ax=ax1, kind="bar", color="lime");
I’m have 2 problem :
1- name of days is not short (Monday : Mon)
2- weekdays are not ordered
Any body can help me to solve this problems ?
Thanks
Advertisement
Answer
One solution can be, Use of pandas.Series.dt.dayofweek
(The day of the week Monday=0, Sunday=6.) Then use .set_xticklabels
to assign the name of day week to the number.
df["dayName"] = pd.to_datetime(df["InvoiceDate"]).dt.dayofweek df.groupby("dayName").size().plot(ax=ax1, kind="bar", color="lime"); ax1.set_xticklabels( ('Mon', 'Tue','Wed', 'Thur', 'Fri', 'Sat', 'Sun') )
Output: (Generate from random input)
[Text(0, 0, 'Mon'), Text(1, 0, 'Tue'), Text(2, 0, 'Wed'), Text(3, 0, 'Thur'), Text(4, 0, 'Fri'), Text(5, 0, 'Sat'), Text(6, 0, 'Sun')]