How to plot the values of a groupby on multiple columns-CodePudding

I have a dataset similar to the following:

import pandas as pd

data = {'Country': ['Spain', 'Italy', 'France', 'Germany', 'Portugal', 'Greece', 'UK', 'Spain', 'Italy', 'France', 'Germany', 'Portugal', 'Greece', 'UK', 'Spain', 'Italy', 'France', 'Germany', 'Portugal', 'Greece', 'UK'],
        'Date': ['Jan 2020', 'Jan 2020', 'Jan 2020', 'Jan 2020', 'Jan 2020', 'Jan 2020', 'Jan 2020', 'Feb 2020', 'Feb 2020', 'Feb 2020', 'Feb 2020', 'Feb 2020', 'Feb 2020', 'Feb 2020', 'Dec 2020', 'Dec 2020', 'Dec 2020', 'Dec 2020', 'Dec 2020', 'Dec 2020', 'Dec 2020'],
        'Sales': [20000, 30000, 10000, 10000, 30000, 10000, 10000, 50000, 40000, 30000, 20000, 30000, 10000, 10000, 60000, 70000, 80000, 10000, 30000, 10000, 10000]}
df = pd.DataFrame(data)

     Country      Date  Sales
0      Spain  Jan 2020  20000
1      Italy  Jan 2020  30000
2     France  Jan 2020  10000
3    Germany  Jan 2020  10000
4   Portugal  Jan 2020  30000
5     Greece  Jan 2020  10000
6         UK  Jan 2020  10000
7      Spain  Feb 2020  50000
8      Italy  Feb 2020  40000
9     France  Feb 2020  30000
10   Germany  Feb 2020  20000
11  Portugal  Feb 2020  30000
12    Greece  Feb 2020  10000
13        UK  Feb 2020  10000
14     Spain  Dec 2020  60000
15     Italy  Dec 2020  70000
16    France  Dec 2020  80000
17   Germany  Dec 2020  10000
18  Portugal  Dec 2020  30000
19    Greece  Dec 2020  10000
20        UK  Dec 2020  10000

I would like to visualize how the Sales varied over the year by Country therefore I would like to show 7 histograms (one for each Country). For each plot, the 'Date' will be on the x-axis and the 'Sales' values on the y-axis. Also, a title to identify the Country is required as well as the x-label, y-label.

I have tried several options found in previous discussions but none of those works with what I want to achieve. I have tried the following:

df.groupby('Country').hist(column='Sales', grid= False, figsize=(2,2))

df['Sales'].hist(grid=True, by=one_year_df['Country'])

df.groupby('Country').hist(grid= False, figsize=(2,2))

df.reset_index().pivot('index','Country','Sales').hist(grid=False, bins=12)

grouped = df.groupby('Country')

ncols=2
nrows = int(np.ceil(grouped.ngroups/ncols))

fig, axes = plt.subplots(nrows=nrows, ncols=ncols, figsize=(12,12), sharey=False)

for (key, ax) in zip(grouped.groups.keys(), axes.flatten()):
    grouped.get_group(key).plot(ax=ax)

ax.legend()
plt.show()

However, none of these options gives me the possibility to set the 'Date' column, also it seems that it is not possible to set the x-axis, y-axis as I wish and as a result, the plots are meaningless.

I have also found another piece of code that seems to consider all the variables but the result still is not as expected:

fig, ax = plt.subplots(figsize=(15,7))
df.groupby(['Country']).sum()['Sales'].plot(ax=ax)
ax.set_xlabel('Date')
ax.set_ylabel('Sales')

Any comments or suggestions are welcome. Thank you.

CodePudding user response：

For each plot, the 'Date' will be on the x-axis and the 'Sales' values on the y-axis. is best shown with a line or bar plot. A histogram is essentially a bar plot (in terms of a visulization).
Convert the 'Date' column to datetime with
- If you plot a bar plot, there will be a crowded mess, because there will be a bar for each row of data.
```
ax = dfp.plot(kind='bar', subplots=True, figsize=(14, 12), layout=(2, 4), rot=0, legend=False)
```