Sectors by Country¶
This notebooks looks at the distribution of sectors for different countries.
In [1]:
countries = (
# North America
("usa-stk", "United States"),
("canada-stk", "Canada"),
("mexico-stk", "Mexico"),
# Asia
("australia-stk", "Australia"),
("japan-stk", "Japan"),
("hongkong-stk", "Hong Kong"),
("singapore-stk", "Singapore"),
# Europe
("spain-stk", "Spain"),
("netherlands-stk", "Netherlands"),
("uk-stk", "United Kingdom"),
("germany-stk", "Germany"),
("france-stk", "France"),
("belgium-stk", "Belgium"),
("swiss-stk", "Switzerland"),
("austria-stk", "Austria"),
("sweden-stk", "Sweden"),
("italy-stk", "Italy"),
)
Get a DataFrame of countries (columns) and sectors (index):
In [2]:
from quantrocket.master import download_master_file
import io
import pandas as pd
all_countries = {}
for universe, country in countries:
f = io.StringIO()
download_master_file(f, universes=universe, fields="ibkr_Sector")
securities = pd.read_csv(f)
all_countries[country] = securities.groupby(securities.ibkr_Sector).Sid.count()
sectors = pd.concat(all_countries).unstack().T
Convert to percentages:
In [3]:
sizes = sectors.sum()
sectors = sectors / sizes
sectors = sectors.round(3) * 100
Transpose:
In [4]:
sectors = sectors.fillna(0).T
Drop minor sectors to simplify plot:
In [5]:
sectors = sectors.drop(["Funds","Government", "Diversified"], axis=1)
Plot:
In [6]:
import matplotlib.pyplot as plt
import seaborn as sns
sns.set(palette=sns.color_palette("hls", n_colors=len(sectors.columns)))
In [7]:
ax = sectors.plot(kind="bar", stacked=True, title="Sectors by country", figsize=(16,6))
ax.set_ylabel("Percentage of total")
plt.legend(loc='center left', bbox_to_anchor=(1, 0.5))
