import pandas as pd

import matplotlib.pyplot as plt

import matplotlib.dates as mdates

import datetime

df = pd.read_csv('dfa-networth-shares.csv')

df['Datex'] = pd.to_datetime(df['Date'].str.replace(':', '-'))

df = df[['Net worth', 'Category', 'Datex']].groupby(['Datex', 'Category']).sum().unstack()

df = df.apply(lambda x: x/x.sum() * 100, axis=1)

df.loc[:, ('Net worth', 'Top10')] = df['Net worth']['Top1'] + df['Net worth']['Next9']

df.loc[:, ('Net worth', 'Bottom90')] = df['Net worth']['Next40'] + df['Net worth']['Bottom50']

df = df.stack()

df = df.reset_index()

plt_df = df[(df['Category'] == 'Top10') | (df['Category'] == 'Bottom90')]

plt.rcParams["font.family"] = "serif"

fig, ax = plt.subplots(figsize=(10, 10), facecolor='white', dpi=300)

COLORS = {

'Top10': '#912d00',

'Bottom90': '#006595'

}

ax.tick_params(axis='y', left=False)

plt.ylim(0, 100)

plt.yticks([25, 50, 75, 100], fontsize=14)

plt.ylabel('Share of total wealth (%)', fontsize=14)

ax.tick_params(axis='x', bottom=False)

plt.xlim(datetime.date(1990,1,1), datetime.date(2022,1,1))

plt.xticks([datetime.date(i,1,1) for i in range(1990, 2021, 10)], fontsize=14)

date_form = mdates.DateFormatter('%Y')

ax.xaxis.set_major_formatter(date_form)

ax.spines['top'].set_visible(False)

ax.spines['right'].set_visible(False)

for k, v in COLORS.items():

year = plt_df.loc[(plt_df['Category'] == k), 'Datex']

worth = plt_df.loc[(plt_df['Category'] == k), 'Net worth']

ax.plot(year, worth, c=v)

ax.plot(year.iloc[-1], worth.iloc[-1], 'o', c=v)

ax.annotate(k + ' percent',

xy=(year.iloc[-1], worth.iloc[-1]),

xytext=(10, -3),

textcoords='offset points',

c=v,

fontsize=12)

plt.title('The U.S. Household Wealth Distribution',

fontsize=28,

loc='left',

x=-0.08, y=1.14)

ax.text(s='The 10 percent wealthiest households possess almost \n70 percent of the total household wealth',

x=-.08, y=1.06,

fontsize=18,

transform=ax.transAxes)

ax.text(s='H | Data: The Federal Reserve',

x=0.5, y=-0.1,

horizontalalignment='center',

fontsize=14,

transform=ax.transAxes)

fig.savefig('wealth', bbox_inches='tight')