How to plot multiple variables with Pandas and Bokeh

Question

I'm new to Pandas and Bokeh; I'd to create a bar plot that shows two different variables next to each other for comparison.

For instance, with the following Pandas data frame, I'd like to see how the amount of Recalled compares to the amount of Recovered for each year.

    year    Recalled    Recovered
0   1994    11.472  10.207
1   1995    11.810  10.326
2   1996    10.632  10.094
3   1997    13.857  12.944
4   1998    13.861  12.588
5   1999    13.375  11.951
6   2000    11.278  nan
7   2001    12.827  nan
8   2002    12.687  nan
9   2003    10.859  nan
10  2004    nan nan
11  2005    11.782  11.047
12  2006    12.089  10.194
13  2007    14.351  13.401
14  2008    14.921  13.886
15  2009    11.759  10.815
16  2010    12.987  11.482
17  2011    13.262  10.730
18  2012    9.980   9.520
19  2013    10.626  9.591
20  2014    12.199  10.270

Answer 1

This changed in the latest version of Bokeh (I guess 0.12.7). This is the new way of doing it.

from bokeh.io import show, output_file
from bokeh.models import ColumnDataSource, FactorRange
from bokeh.plotting import figure

output_file("bars.html")

fruits = ['Apples', 'Pears', 'Nectarines', 'Plums', 'Grapes', 'Strawberries']
years = ['2015', '2016', '2017']

data = {'fruits' : fruits,
        '2015'   : [2, 1, 4, 3, 2, 4],
        '2016'   : [5, 3, 3, 2, 4, 6],
        '2017'   : [3, 2, 4, 4, 5, 3]}

# this creates [ ("Apples", "2015"), ("Apples", "2016"), ("Apples", "2017"), ("Pears", "2015), ... ]
x = [ (fruit, year) for fruit in fruits for year in years ]
counts = sum(zip(data['2015'], data['2016'], data['2017']), ()) # like an hstack

source = ColumnDataSource(data=dict(x=x, counts=counts))

p = figure(x_range=FactorRange(*x), plot_height=250, title="Fruit Counts by Year",
           toolbar_location=None, tools="")

p.vbar(x='x', top='counts', width=0.9, source=source)

p.y_range.start = 0
p.x_range.range_padding = 0.1
p.xaxis.major_label_orientation = 1
p.xgrid.grid_line_color = None

show(p)

Answer 2

Just pandas, no bokeh (copy the data to the clipboard before running):

import pandas, seaborn
DF = pandas.read_clipboard()
DF.plot.bar(x='year')

Answer 3

You can use grouping in the Bokeh high-level bar chart if you first melt your Pandas dataframe.

import pandas as pd
from bokeh.plotting import figure, show

# Use output_notebook if you are using an IPython or Jupyter notebook
from bokeh.io import output_notebook
output_notebook()

# Get your data into the dataframe
df = pd.read_csv("data.csv")

# Create a "melted" version of your dataframe
melted_df = pd.melt(df, id_vars=['year'], value_vars=['Recalled', 'Recovered'])

melted_df.head()

Here's the format of your melted dataframe:

+---+------+----------+--------+
|   | year | variable | value  |
+---+------+----------+--------+
| 0 | 1994 | Recalled | 11.472 |
| 1 | 1995 | Recalled |  11.81 |
| 2 | 1996 | Recalled | 10.632 |
| 3 | 1997 | Recalled | 13.857 |
| 4 | 1998 | Recalled | 13.861 |
+---+------+----------+--------+

Then just use your melted dataframe as your data in the Bokeh bar chart:

p = Bar(melted_df, label="year", values="value", group="variable", legend="top_left", ylabel='Values')
show(p)

Answer 4

This is my answer just using matplotlib and numpy.

My code seem to be very lengthy comparing with the accepted answer. Hence, if someone can help me to improve it, that will be great!

Code here:

## Reading the data
year = np.arange(1994,2015,1)
type = ["Recalled", "Recovered"]
value_1 = [11.472, 11.81, 10.632, 13.857, 13.861, 13.375, 11.278, 12.827, 12.687, 10.859, np.nan, 11.782, 12.089,\
       14.351, 14.921, 11.759, 12.987, 13.262, 9.98, 10.626, 12.199]
value_2 = [10.207, 10.326, 10.094, 12.944, 12.588, 11.951, np.nan, np.nan, np.nan, np.nan, np.nan, 11.047, 10.194,\
       13.401, 13.886, 10.815, 11.482, 10.73, 9.52, 9.591, 10.27] 



## Changing data into ndarray format
dpoints = np.array([type[0], year[0], value_1[0]])
conditions =  np.unique(dpoints[:,0])
categories =  np.unique(dpoints[:,1]).tolist()

for i in range(1,len(year),1):
    dpoints = np.vstack([dpoints,np.array([type[0],year[i],value_1[i]])])
for i in range(0,len(year),1):
    dpoints = np.vstack([dpoints,np.array([type[1],year[i],value_2[i]])])

# Plot it!
fig = plt.figure(figsize=(16,5))
ax = plt.subplot()
#the space between each set of bars
space = 0.2
n = len(conditions)
width = (1 - space) / (len(conditions))

# Create a set of bars at each position
for i,cond in enumerate(conditions):
    indeces = range(1, len(categories)+1)
    vals = dpoints[dpoints[:,0] == cond][:,2].astype(np.float)
    pos = [j - (1 - space) / 2. + i * width for j in indeces]

    ax.bar(pos, vals, width=width, label =cond,lw = 0,color=  ["blue","r"][i],alpha = 0.6)


    # Set the x-axis tick labels to be equal to the categories
    ax.set_xticks(indeces)
    ax.set_xticklabels(categories)
    ax.set_xlim(0,22)
    plt.setp(plt.xticks()[1], rotation=0,fontsize = 12)
    ax.set_ylabel('Variables',fontsize =15)
    ax.set_xlabel("Year",fontsize =15)


    # Add a legend
    handles, labels = ax.get_legend_handles_labels()
    ax.legend(handles[::-1], labels[::-1], loc='upper left',frameon =False,fontsize =12)    

Figure here