Plot time series with colorbar in pandas + matplotlib

Question

I'm trying to plot a colorbar below this chart, where the color depends on when each of the time series starts:

The code generated to create the plot is this:

import pandas as pd
import numpy as np

import matplotlib.pyplot as plt

import seaborn as sns
sns.set()

def partial_cum_returns(start, cum_returns):
    return cum_returns.loc[start:].div(cum_returns.loc[start])

index = pd.DatetimeIndex(pd.date_range('20170101', '20190101', freq='W'))
np.random.seed(5)
returns = pd.Series(np.exp(np.random.normal(loc=0, scale=0.05, size=len(index))), index=index)
cum_returns = returns.cumprod()
df = pd.DataFrame(index=index)
for date in index:
    df[date] = partial_cum_returns(date, cum_returns)

df.plot(legend=False, colormap='viridis');
plt.colorbar();

But when executing this error appears:

RuntimeError: No mappable was found to use for colorbar creation. First define a mappable such as an image (with imshow) or a contour set (with contourf).

I've tried to add the colorbar in different ways, like the fig, ax = plt.figure()... one, but I couldn't make it work so far. Any ideas? Thanks!

ALollz · Accepted Answer · 2018-07-03 19:22:45Z

7

The first point is that you need to create a ScalarMappable for your colorbar. You need to define the colormap, which in your case is 'viridis' and specify the maximum and minimum of the values you want for the colorbar. Then because it uses numeric time values you want to reformat those.

import matplotlib.pyplot as plt
import pandas as pd

# Define your mappable for colorbar creation
sm = plt.cm.ScalarMappable(cmap='viridis', 
                           norm=plt.Normalize(vmin=df.index.min().value,
                                              vmax=df.index.max().value))
sm._A = []  

df.plot(legend=False, colormap='viridis', figsize=(12,7));

cbar = plt.colorbar(sm);
# Change the numeric ticks into ones that match the x-axis
cbar.ax.set_yticklabels(pd.to_datetime(cbar.get_ticks()).strftime(date_format='%b %Y'))

edited Jul 3, 2018 at 19:22

answered Jul 3, 2018 at 19:14

ALollz

59.7k7 gold badges73 silver badges97 bronze badges

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

4 Comments

Xoel Over a year ago

Thanks! I used this code to make the colorbar horizontal: cbar = plt.colorbar(sm, orientation='horizontal', pad=0.1, aspect=40, shrink=0.9);

Xoel Over a year ago

Also, what does the line sm._A = [] do in your code?

ALollz Over a year ago

@XoelLópezBarata As far as I know, mappables need an array. Initializing a scalar mappable in the way I did initializes it's array to None, which leads to a TypeError: You must first set_array for mappable. That line simply sets the array to the empty list. Though I'm not sure what you would use it for. Someone with more expertise with color bars would be able to answer that.

vicemagui Over a year ago

@Xoel did you find a solution? I'm having the same problem now

Collectives™ on Stack Overflow

Plot time series with colorbar in pandas + matplotlib

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related