Create "The Economist" style graphs from python

Question

Using python and marplotlib and a tool like seaborn, I'd like to create a graph like this one from The Economist (because I think the style is great.)

It's a time series graph and the key things I'd like to reproduce are the horizontal grid lines with labels matched with a lower horizontal axis with tick marks. The different colored labels at either end of the grid lines would be a bonus, with the corresponding titles (left and right justified). The annotations would be a double bonus.

I tried to make something like it using seaborn, but could not get to the first step.

Answer 1

Not perfect (I've haven't had long to play with it), but to give you an idea of the kind of Matplotlib methods you need to use to customize a plot the way you want, there's some code below.

Note that to fine-tune a plot like this it is hard to keep the content and presentation separate (You may have to manually set tick labels and the like, so it won't work automatically if you change the data). The Economist's graphics people obviously do this because they seem to have got the top left hand tick label wrong (280 should be 260).

# -*- coding: utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import matplotlib.ticker as ticker
from datetime import datetime

# Load in some sample data
bond_yields = np.loadtxt('bond_yields.txt',
                         converters={0: mdates.strpdate2num('%Y-%m-%d')},
                         dtype = {'names': ('date', 'bond_yield'),
                                  'formats': (datetime, float)})
bank_deposits = np.loadtxt('bank_deposits.txt',
                         converters={0: mdates.strpdate2num('%Y-%m-%d')},
                         dtype = {'names': ('date', 'bank_deposits'),
                                  'formats': (datetime, float)})

# Bond yields line is in light blue, bank deposits line in dark red:
bond_yield_color = (0.424, 0.153, 0.090)
bank_deposits_color = (0.255, 0.627, 0.843)

# Set up a figure, and twin the x-axis so we can have two different y-axes
fig = plt.figure(figsize=(8, 4), frameon=False, facecolor='white')
ax1 = fig.add_subplot(111)
ax2 = ax1.twinx()
# Make sure the gridlines don't end up on top of the plotted data
ax1.set_axisbelow(True)
ax2.set_axisbelow(True)

# The light gray, horizontal gridlines
ax1.yaxis.grid(True, color='0.65', ls='-', lw=1.5, zorder=0)

# Plot the data
l1, = ax1.plot(bank_deposits['date'], bank_deposits['bank_deposits'],
         c=bank_deposits_color, lw=3.5)
l2, = ax2.plot(bond_yields['date'], bond_yields['bond_yield'],
         c=bond_yield_color, lw=3.5)

# Set the y-tick ranges: chosen so that ax2 labels will match the ax1 gridlines
ax1.set_yticks(range(120,280,20))
ax2.set_yticks(range(0, 40, 5))

# Turn off spines left, top, bottom and right (do it twice because of the twinning)
ax1.spines['left'].set_visible(False)
ax1.spines['right'].set_visible(False)
ax1.spines['top'].set_visible(False)
ax2.spines['left'].set_visible(False)
ax2.spines['right'].set_visible(False)
ax2.spines['top'].set_visible(False)
ax1.spines['bottom'].set_visible(False)
ax2.spines['bottom'].set_visible(False)
# We do want ticks on the bottom x-axis only
ax1.xaxis.set_ticks_position('bottom')
ax2.xaxis.set_ticks_position('bottom')

# Remove ticks from the y-axes
ax1.tick_params(axis='y', length=0)
ax2.tick_params(axis='y', length=0)

# Set tick-labels for the two y-axes in the appropriate colors
for tick_label in ax1.yaxis.get_ticklabels():
    tick_label.set_fontsize(12)
    tick_label.set_color(bank_deposits_color)
for tick_label in ax2.yaxis.get_ticklabels():
    tick_label.set_fontsize(12)
    tick_label.set_color(bond_yield_color)

# Set the x-axis tick marks to two-digit years
ax1.xaxis.set_major_locator(mdates.YearLocator())
ax1.xaxis.set_major_formatter(mdates.DateFormatter('%y'))

# Tweak the x-axis tick label sizes
for tick in ax1.xaxis.get_major_ticks():
    tick.label.set_fontsize(12)
    tick.label.set_horizontalalignment('center')

# Lengthen the bottom x-ticks and set them to dark gray
ax1.tick_params(direction='in', axis='x', length=7, color='0.1')

# Add the line legends as annotations
ax1.annotate(u'private-sector bank deposits, €bn', xy=(0.09, 0.95),
             xycoords='figure fraction', size=12, color=bank_deposits_color,
             fontstyle='italic')
ax2.annotate(u'ten-year government bond yield, %', xy=(0.6, 0.95),
             xycoords='figure fraction', size=12, color=bond_yield_color,
             fontstyle='italic')

# Add an annotation at the date of the first bail-out. relpos=(0,0) ensures
# that the label lines up on the right of a vertical line
first_bailout_date = datetime.strptime('2010-05-02', '%Y-%m-%d')
xpos = mdates.date2num(first_bailout_date)
ax1.annotate(u'FIRST BAIL-OUT', xy=(xpos, 120), xytext=(xpos, 250), color='r',
             arrowprops=dict(arrowstyle='-', edgecolor='r', ls='dashed',
             relpos=(0,0)), fontsize=9, fontstyle='italic')

fig.savefig('fig.png', facecolor=fig.get_facecolor(), edgecolor='none')