CODEWITHSUNDEEP
pythonpandasmatplotlibbar-chart
lanery
lanery

Reputation: 5364

How to deal with annoying gap in pandas bar plot

I would like to fix the gap between 2012 and 2013 in the below bar plot.

enter image description here

My dataframe is

In [30]: df
Out[30]:
            Pre-Release  Post-Release
FinishDate
2008                1.0           0.0
2009               18.0           0.0
2010               96.0           0.0
2011              161.0           0.0
2012              157.0           0.0
2013                0.0         139.0
2014                0.0         155.0
2015                0.0         150.0
2016                0.0          91.0
2017                0.0          15.0

and I'm using df.plot(kind='bar', width=1) to plot.

Upvotes: 6

Views: 2253

Answers (3)

tenhjo
tenhjo

Reputation: 4537

You plot two data sets for each year. So at each node on the x-axis two bars for both of the data sets are plotted. You can't see them because the values are zero. I think the keyword stacked=True might work. This stacks both data sets vertically and no gap shows.

Upvotes: 1

j-i-l
j-i-l

Reputation: 10957

The gap is actually often the wanted behaviour as you are effectively plotting two histograms here.

In this case, however, the reported values seem to be exclusive, so there is no need for plotting the histograms side-by-side, instead just plot one and then the other.

Here is a minimal example that should get you there:

import matplotlib.pyplot as plt
import numpy as np
import pandas

someDF = pandas.DataFrame({'1':np.random.uniform(size=20)*10, '2':np.random.uniform(size=20)*10})

cut = 10
fig, ax = plt.subplots(figsize=(8,3))

first = someDF['1'][someDF.index >= cut]
second = someDF['2'][someDF.index < cut]

ax.bar(left=first.index, height=first, align='center', color='blue')
ax.bar(left=second.index, height=second, align='center', color='red')
plt.show()

Output looks then something like: enter image description here

Upvotes: 1

Sebastian Wozny
Sebastian Wozny

Reputation: 17506

There is no actual "Gap" in your graph: Pandas is just reserving space to plot two different bars next to each other. Take this piece of code to illustrate:

from io import StringIO
import pandas as pd
TESTDATA=StringIO("""2008                1.0           0.0
2009               18.0           5.0
2010               96.0           0.0
2011              161.0           0.0
2012              157.0           0.0
2013                0.0         139.0
2014                0.0         155.0
2015                0.0         150.0
2016                0.0          91.0
2017                0.0          15.0""")
df=pd.read_csv(TESTDATA,delim_whitespace=True,index_col=0)
df.plot(kind='bar')

Two bars next to each other

But you don't actually ever need to print two bars next to each other, so rather than plotting the dataframe, you can plot two series into the same figure:

ax=df['1.0'].plot(kind='bar')
df['0.0'].plot(kind='bar',ax=ax,color='orange')

enter image description here

Alternatively just use:

df.plot(kind='bar', stacked=True)

Which gives you the same results in this case.

Upvotes: 4

Related Questions