I have a dataframe with a varying number of columns. With positive and negative values. Is it possible to make a stacked bar chart with all of these columns in the dataframe?
A B C
0 34 34 -12
1 34 223 -12
2 34 56 -12
3 34 86 -12
4 56 86 -12
5 56 43 -12
6 78 34 -12
(Updated answer for newer versions of plotly)
Using px.bar will give you a stacked bar chart directly. If your bar chart for some reason is not stacked, try:
fig.update_layout(barmode='stack')
import pandas as pd
import plotly.express as px
# data
df = pd.DataFrame({'A': {0: 34, 1: 34, 2: 34, 3: 34, 4: 56, 5: 56, 6: 78},
'B': {0: 34, 1: 223, 2: 56, 3: 86, 4: 86, 5: 43, 6: 34},
'C': {0: -12, 1: -12, 2: -12, 3: -12, 4: -12, 5: -12, 6: -12}})
colors = px.colors.qualitative.T10
# plotly
fig = px.bar(df,
x = df.index,
y = [c for c in df.columns],
template = 'plotly_dark',
color_discrete_sequence = colors,
title = 'Stacked bar chart using px.bar()',
)
fig.show()
Old answer:
Here's a stacked bar chart using plotly offline in a Jupyter Notebook. Just make sure you have the packages listed under imports installed. df = cf.datagen.lines() produces a pandas dataframe with sample data from cufflinks.
Plotly stacked bar chart:
Reproducible snippet:
# imports
import plotly
import cufflinks as cf
from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot
import pandas as pd
import numpy as np
# setup
init_notebook_mode(connected=True)
np.random.seed(123)
cf.set_config_file(theme='pearl')
# qtconsole for debugging
#%qtconsole --style vim #src# https://qtconsole.readthedocs.io/en/stable/
# Random data using cufflinks
df = cf.datagen.lines()
df = df[['UUN.XY', 'MJF.XV', 'XBB.AO']].head(50)
df=np.abs(df)
# make figure
fig = df.iplot(asFigure=True, kind='bar', barmode = 'stack',
xTitle='Dates',yTitle='Returns',title='Returns')
# plot figure
iplot(fig)
Edit - Test with negative values
Here's the same thing run without the line df=np.abs(df)
y is 3, whereas the length of previous arguments ['x'] is 7It is now possible to make stacked bar plot directly with pandas by using plotly as plotting backend.
pd.options.plotting.backend = 'plotly'
fig = df.plot(kind='bar')
