Plotting percentages of occurrences in a stacked barchart

Question

I have trying to create a stacked bar chart that will show the percentage which each item occurred over a given year.

The problem is that when I plot these values, not all the bar's show. Seems like some of the bar's are being masked by the bars that are showing.

This is the relevant code:

barWidth = 0.85
plt.bar(list(yearly_timeline.index),yearly_timeline.values, color='#a3acff',edgecolor='white',width=barWidth)
plt.bar(list(yearly_links.index),yearly_links.values, color='#FFD700',edgecolor='white',width=barWidth)
plt.bar(list(yearly_status.index),yearly_status.values, color='#b5ffb9',edgecolor='white',width=barWidth)
plt.bar(list(yearly_posts.index),yearly_posts.values,color='#f9bc86',edgecolor='white',width=barWidth)
plt.bar(list(yearly_shared.index),yearly_shared.values,color='#f9bc86',edgecolor='white',width=barWidth)
plt.xticks(list(yearly_links.index))
fig = plt.gcf()
fig.set_size_inches(20,10)
plt.tick_params(labelsize=20)
plt.show()

This is a sample of the datasets I am plotting:

#yearly posts
year
2009     4.907975
2010    11.656442
2013    11.656442
2014    24.539877
2015     7.975460
2016    12.269939
2017    16.564417
2018    10.429448
dtype: float64

#yearly shared
year
2010     1.273885
2011     0.636943
2012     9.554140
2013    29.936306
2014    28.025478
2015    15.923567
2016     7.643312
2017     4.458599
2018     2.547771
dtype: float64

#yearly timeline
year
2010     4.059041
2011    18.450185
2012    18.819188
2013    12.915129
2014    25.830258
2015    16.236162
2016     2.214022
2017     1.107011
2018     0.369004
dtype: float64

#yearly status
year
2009     6.916192
2010     6.997559
2011    15.296989
2012    22.294548
2013    19.528072
2014    13.913751
2015    10.740439
2016     1.790073
2017     1.464605
2018     1.057771
dtype: float64

#yearly links
year
2009     0.655738
2010     0.218579
2011     8.196721
2012     8.524590
2013     1.530055
2014     7.103825
2015    26.338798
2016    17.595628
2017    25.027322
2018     4.808743
dtype: float64

Read [this](https://stackoverflow.com/questions/44309507/stacked-bar-plot-using-matplotlib) — Sheldore, Jan 14 '19 at 20:08
@Bazingaa Thanks, but the adding a bottom each dataset that is sum of all the datasets that came before returns a 'ValueError: shape mismatch: objects cannot be broadcast to a single shape' error message ' — Emm, Jan 14 '19 at 20:49

score 0 · Answer 1 · answered Jan 14 '19 at 21:42

In your case, you could simplify your code by aggregating all your data in a single DataFrame (I assume they are currently individual Series):

create dummy data

my_names = ['timeline','links','status','posts','shared']
my_series = [pd.Series(data=np.random.random(size=(9,)), index=range(2010,2019), name=n) for n in my_names]

convert list of Series into a DataFrame:

df = pd.DataFrame(my_series).T
display(df)

      timeline     links    status     posts    shared
2010  0.534663  0.107604  0.265774  0.849307  0.149886
2011  0.064561  0.354329  0.557265  0.297695  0.563122
2012  0.646828  0.011643  0.608695  0.493709  0.337949
2013  0.170792  0.083039  0.866962  0.278223  0.501074
2014  0.386262  0.979529  0.972009  0.333049  0.505644
2015  0.764539  0.223265  0.365314  0.712091  0.757626
2016  0.012084  0.700645  0.118666  0.118811  0.332993
2017  0.407492  0.480495  0.399464  0.613331  0.655171
2018  0.072698  0.262846  0.763811  0.783575  0.859755

The easy way, using pandas plot command:

df.plot(kind='bar', stacked=True, width=0.85)

or using matplotlib directly, to give more flexibility:

fig, ax = plt.subplots()
for i,col in enumerate(df.columns):
    ax.bar(df.index, height=df[col], bottom=df.iloc[:,:i].sum(axis=1), edgecolor="white", width=0.85)
ax.set_xticks(df.index)

Plotting percentages of occurrences in a stacked barchart

1 Answers1