# 2. could newer data have come already?
# 3. [+] let's parse the new data
# 4. [-] commit to repo
# 5. upload to database
# 6. create a PDF handout or other assets
# plotting
import matplotlib.pyplot as plt
from access import get_dataframe
dfa, dfq, dfm = (get_dataframe(freq) for freq in 'aqm')
start_year = 2016
renamer = {'INDPRO_yoy': ('Промышленное производство', 'темп прироста за 12 мес., %'),
'GDP_yoy': ('Валовый внутренний продукт', 'темп прироста за 12 мес., %'),
'RETAIL_SALES_yoy': ('Розничные продажи', 'темп прироста за 12 мес., %'),
'CPI_rog': ('Индекс потребительских цен', 'к пред.периоду, %'),
'WAGE_REAL_rog':('Реальная заработная плата', 'к пред.периоду, %'),
'AGROPROD_yoy':('Сельскохозяйственное производство', 'темп прироста за 12 мес., %')
}
fig, axes = plt.subplots(nrows=2, ncols=2, figsize=(4*3.5, 3*3.5))
ax = df.plot(x_compat=True)
# set ticks
ax.set_xticks(minor_ticks, minor=True)
ax.xaxis.set_ticks(major_ticks, minor=False)
ax.xaxis.set_major_formatter(dates.DateFormatter('%Y'))
plt.gcf().autofmt_xdate(rotation=0, ha="center")
# other formatting
plt.legend(loc="lower left")
ax.set_title(title, loc='left', fontdict={'fontsize': 11})
return ax
def plot_long(df, title, start=2005, end=2020, left_offset=1):
"""Plot starting 2005."""
df = df[df.index >= str(start)]
minor_ticks = pd.date_range(str(start - left_offset), str(end), freq='YS')
major_ticks = pd.date_range(str(start), str(end), freq='5YS')
return plot(df, title, minor_ticks, major_ticks)
def save(filename):
plt.savefig(filename)
if __name__ == '_main__':
import access
dfa, dfq, dfm = (access.get_dataframe(freq) for freq in 'aqm')
df = dfm.DWELLINGS_CONSTRUCTION_mln_m2
ax = plot_long(df, 'Ввод жилья, млн.кв.м')
plt.show()
def many_plots_per_page(df, nrows, ncols, figsize=A4_SIZE_PORTRAIT, title_font_size=TITLE_FONT_SIZE):
page_vars = df.columns
# The following command uses the built-in Pandas mechanism for placing subplots on a page.
# It automatically increases spacing between subplots and rotates axis ticks if they
# take up too much space. However, this mechanism is broken in Pandas < 0.17.
# See: https://github.com/pydata/pandas/issues/11536
# It also cannot handle multiple variables per subplot, so if we want that, we'll have to
# replicate parts of the Pandas implementation or write our own.
axes = df.plot(subplots=True, layout=(nrows, ncols), legend=None, figsize=figsize)
# Now removing axis labels and adding plot titles.
for i, axes_row in enumerate(axes):
for j, ax in enumerate(axes_row):
var_idx = i * ncols + j
if var_idx >= len(page_vars):
# We're at the end of the last page, which is not filled completely.
break
ax.set_title(page_vars[var_idx], fontsize=title_font_size)
format_ax(ax)
return axes
if __name__== '__main__':
import access
dfm = access.get_dataframe('m')
save_plots_as_pdf(dfm, 'plt.pdf')