python plot multiple bar ranges with dates

Question

I have this two dataframe:

x1=[{"dates":'2018-01-31',"rev":-2},
{"dates":'2018-02-28',"rev":-5},
{"dates":'2018-03-31',"rev":-7},
{"dates":'2018-04-30',"rev":-8},
{"dates":'2018-05-31',"rev":-9},
{"dates":'2018-06-30',"rev":-2},
{"dates":'2018-07-31',"rev":1},
{"dates":'2018-08-31',"rev":2},
{"dates":'2018-09-30',"rev":3},
{"dates":'2018-10-31',"rev":4},
{"dates":'2018-11-30',"rev":4},
{"dates":'2018-12-31',"rev":5}]


x2=[{"dates":'2018-01-31',"rev":-5},
{"dates":'2018-02-28',"rev":-9},
{"dates":'2018-03-31',"rev":-9},
{"dates":'2018-04-30',"rev":-6},
{"dates":'2018-05-31',"rev":-1},
{"dates":'2018-06-30',"rev":-2},
{"dates":'2018-07-31',"rev":-14},
{"dates":'2018-08-31',"rev":2},
{"dates":'2018-09-30',"rev":3},
{"dates":'2018-10-31',"rev":4},
{"dates":'2018-11-30',"rev":-4},
{"dates":'2018-12-31',"rev":5}]


df1=pd.DataFrame(x1)
df1["dates"]=pd.to_datetime(df1.dates)
df2=pd.DataFrame(x2)
df2["dates"]=pd.to_datetime(df2.dates)

(Thanks Flavio Moraes )

I would like to plot then as a bar plot. The matplotlib guide suggests to do something like the following:

width = 0.35  
fig, ax = plt.subplots()
rects1 = ax.bar(x - width/2, y1, width, label='y1')
rects2 = ax.bar(x + width/2, y2, width, label='y2')

However, this seems to not working with dates. Indeed I have tried:

ax = plt.axes()
widthb = 10
rects1 = ax.bar(df1.dates - widthb/2, df1.rev,color='blue', width=widthb)
rects2 = ax.bar(df2.dates + widthb/2, df2.rev,color='red',  width=widthb)
ax.xaxis_date()

As expected, I get the following error:

unsupported operand type(s) for -: 'DatetimeIndex' and 'float'

Do you have any suggestion to overcome this problem and simultaneously to have a beautiful plot?

Thanks

Why not just use pandas's plotting methods, i.e. DataFrame.plot.bar. — BigBen
– BigBen, Commented Nov 16, 2020 at 21:53
You can create the axes and pass it to pandas though, let pandas do the lion's share of the plotting, then customize the output to your heart's content w/ matplotlib. — BigBen
– BigBen, Commented Nov 16, 2020 at 22:08
it's quite difficult to do that unless you do without some important feature of matplolib. For example I have named a column as "ex_ax". Due to the fact that I would like to plot according to latex style (matplotlib.rc('text', usetex=True)), pandas was not able to handle it in plot.bar . — diedro
– diedro, Commented Nov 17, 2020 at 23:35

Quang Hoang · Accepted Answer · 2020-11-16 22:09:09Z

1

Let's try merge and plot:

ax = df1.merge(df2, on='dates', how='outer').plot.bar(x='dates')
# other format with `ax`
ax.xaxis_date()

Output:

answered Nov 16, 2020 at 22:09

Quang Hoang

151k11 gold badges64 silver badges86 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

diedro Over a year ago

thanks a lot. I will use this solution, although I prefer to plot with ax and matplolib features.

Quang Hoang Over a year ago

@diedro ax is a Matplotlib axis instance. You can do everything you would do with other axis instance. ax.xaxis_date() is an example.

Collectives™ on Stack Overflow

python plot multiple bar ranges with dates

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related