"""Download and truncate Rosstat corporate dataset."""
from boo import download, build, read_dataframe, files
print("Please be prepared: "
"download and build operations "
"can take long time!")
year = 2012
# Download raw file from Rosstat
try:
download(year)
except FileExistsError:
print("Raw file already downloaded")
# Select fewer columns and assign short column names
# Will save to new file
try:
build(year)
except FileExistsError:
print("Work file already created")
# Read data as dataframe
df = read_dataframe(year)
print(year, "dataset:", df.shape[0], "rows and", df.shape[1], "columns")
print("File locations:", files(year))
from boo import download, read_dataframe download(2012) df = read_dataframe(2012) print(df.head())
import matplotlib.pyplot as plt
from boo import read_dataframe
import pick
try:
df
except NameError:
df = read_dataframe(2017)
try:
df0
except NameError:
df0 = pick.filter0(df)
try:
bs
except NameError:
bs = pick.nlargest(df0, 'sales', 500)
def ab(t, n=20):
return df[df.ok1 == t].head(n)
def bln(x):
return str(round(x / 10**6, 0)).rjust(5)
# print as tables
for b in bs.itertuples():
print(b.inn, str(b.ok1).rjust(2), bln(b.sales), bln(b.cf_oper),
cols = numeric_columns(df)
zf = df.copy()
zf.loc[:, cols] = zf.loc[:, cols].divide(divide_by).round(digits)
return zf
# save as CSV and Excel
def locate(filename):
return os.path.join("assets", filename)
if __name__ == "__main__":
must_overwrite = True
boo.download(2018)
source_df = boo.read_dataframe(2018)
print("Finished reading file, querying...")
# Has some profit or loss, but not exactly zero thousand RUB,
# (protects from ghost firms)
ix = source_df.profit_before_tax != 0
# Not a financial firm
ix = ix & (~source_df.ok1.isin([64, 65]))
# Gazprom will be on top of list
df = source_df[ix].sort_values("ta", ascending=False).dropna()
if must_overwrite:
print("Saving files...")
df1 = change_unit(df, divide_by=1_000_000, digits=3).query("ta>1")
def make_df0(year):
df = read_dataframe(year)
n_dups = inn_duplicates(df)
print(f"Cleared {n_dups} duplicates from dataset. All rows unique.")
return df.set_index('inn')
return df[(df.ok1 == ok1) & (df.ok2 == ok2)]
def sales_df(df):
return sort_sales(df)[SMALL_SHOW]
def ta_df(df):
return sort_sales(df)[SMALL_SHOW]
if __name__ == "__main__":
try:
df
except NameError:
df = read_dataframe(2017).set_index('inn')
# 1. Показать крупнейшие компании по продажам и объему активов
# ============================================================
base_df = base_report(df)
sf = sort_sales(base_df)[SMALL_SHOW]
af = sort_ta(base_df)[SMALL_SHOW]
n = 5
print("\nКрупнейшие компании по выручке:")
print(sf.head(n))
print("\nКрупнейшие компании по активам:")
print(af.head(n))
# - не вычищены финановые компании, у которых большая выручка
# - есть компании-призраки