Pandas Dataframe plot not showing dates when matplotlib.dates used - python-3.x

Question

I have the following code that plots COVID-19 confirmed cases country-wise against some dates.
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
df = pd.DataFrame({'Countries': ['Australia', 'India', 'UAE', 'UK'],
'3/1/20': [ 27, 3, 21, 36],
'3/2/20': [ 30, 5, 21, 40],
'3/3/20': [ 39, 5, 27, 51],
'3/4/20': [ 52, 28, 27, 86],
},
index = [0, 1, 2, 3])
print('Datframe:\n')
print(df)
dft=df.T
print('\n Transposed data:\n')
print(dft)
print(dft.columns)
dft.columns=dft.iloc[0]
dft=dft[1:]
print('\n Final data:\n')
print(dft)
dft.plot.bar(align='center')
# Set date ticks with 2-day interval
plt.gca().xaxis.set_major_locator(mdates.DayLocator(interval=2))
# Change date format
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%d-%m-%Y'))
''' Note: If I comment above two lines, I get back x-axis ticks. '''
# Autoformatting dates ticks
plt.gcf().autofmt_xdate()
plt.title('COVID-19 confirmed cases')
plt.show()
Here I intended to show the dates on the x-axis ticks with 2-day intervals and get the dates formatted in a different style. However, in the plot, I don't get any ticks and labels on the x-axis as shown in the figure below.
However, when I comment out the instructions with matplotlib.dates, I get back the x-ticks and labels.
Can this be explained and fixed in a simple way? Also, can we get the same result using fig, ax = plt.subplots()?

Answer 1

You were almost there. All you need to do is to restructure your dataframe. index the date. One way to do this is as follows;
Data
df = pd.DataFrame({'Countries': ['Australia', 'India', 'UAE', 'UK'],
'3/1/20': [ 27, 3, 21, 36],
'3/2/20': [ 30, 5, 21, 40],
'3/3/20': [ 39, 5, 27, 51],
'3/4/20': [ 52, 28, 27, 86],
},
index = [0, 1, 2, 3])
df2=df.set_index('Countries').T.unstack().reset_index()
df2#.plot(kind='bar')
df2.columns=['Countries','Date','Count']
df2['Date']=pd.to_datetime(df2['Date'])
df2.dtypes
Coarce date to datetime
df2.set_index('Date', inplace=True)
Groupby date, countries and unstack before plotting
df2.groupby([df2.index.date,df2['Countries']])['Count'].sum().unstack().plot.bar()
Outcome