I am trying to create a plot like the one below. It seems like a bar plot, but I haven’t found a way to control the individual linewidths to reproduce such a plot. I’ll appreciate any help or insight, thanks in advance.
source: The AI Economist: Improving Equality and Productivity with AI-Driven Tax Policies
This type of plot is implemented in the development in the github install from source version as matplotlib.axes.Axes.stairs — Matplotlib 3.3.4.post2631+g645b146eb documentation
I had a similar problem and wanted to get rid of the vertical lines. Here is a short script to create a figure similar to the one you posted. It is not perfect and slightly messy, but I hope it helps you 
import numpy as np
from matplotlib import pyplot as plt
from matplotlib import legend as mlegend
# fix random seed for reproducibility
np.random.seed(42)
# generate data
nlines = 3
labels = ['US Federal', 'Saez Formula', 'Al Economist']
nevents = 7
x = np.cumsum(
[0, *np.random.randint(low=10, high=50, size=nevents)]
)
data = np.random.randint(
low=20,
high=80,
size=(nlines, nevents),
)
# plotting data
fig, axs = plt.subplots(
3,
1,
sharey=True,
sharex=True,
figsize=(8, 4),
)
for idx, (y, ax, label) in enumerate(zip(data, axs.flatten(), labels)):
color = f'C{idx}',
# add horizontal lines
ax.hlines(
y,
xmin=x[:-1],
xmax=x[1:],
label=label,
linewidth=3,
color=color,
)
# fill under line
ax.fill_between(
x,
[*y, y[-1]], # dublicate last value to draw last step as well
step='post',
facecolor=color,
edgecolor=None, # hide edge
alpha=0.5,
)
# hide ticks and axes
ax.tick_params(
axis='both',
which='both',
length=0, # remove ticks
pad=6, # enlarge padding
)
ax.set_frame_on(False)
# set limits, this can be achieved as well via rcParams
ax.set_xlim([np.min(x), np.max(x)])
ax.set_ylim([0, ax.get_ylim()[1]])
ax.set_xticks(x[1:-1])
# plot shared legend
axs[0].legend(
# get handles and labels of multiple axes
*mlegend._get_legend_handles_labels(axs),
bbox_to_anchor=(0.5, 1),
loc='lower center',
ncol=nlines,
frameon=False,
)
# add x labels
axs[-1].set_xlabel('Income (Coin)', weight='bold')
# shared y label hack
# add axes of size of whole canvase
ax = fig.add_subplot(111, frameon=False)
ax.set_ylabel('Marginal Tax Rate (%)', weight='bold')
# remove ticks to get no artificial padding
ax.set_yticks([])
ax.set_xticks([])
fig.align_labels()
plt.show()
Which creates
