def plot_rank(title='Inzidenz',
func=weekly_r,
states=DE_STATE_NAMES,
population=DE_STATE_POPULATION):
global cached_dfs
fig, axes = plt.subplots(nrows=4, ncols=4, sharex=True, sharey=True)
de = None
for iso, state in states.items():
if iso not in cached_dfs:
cached_dfs[iso] = entorb.to_dataframe(iso)
if de is None:
de = func(cached_dfs[iso], population=population[state])
de.drop(columns=[c for c in de.columns if c != 'Cases'])
de.rename(columns={'Cases': state}, inplace=True)
else:
de[state] = func(cached_dfs[iso],
population=population[state])['Cases']
rnk = de.rank(axis=1).rolling('30D').mean()
ax_idx = 0
for iso, state in states.items():
rnk[state].plot(kind='line',
ax=axes.flat[ax_idx],
title=state,
legend=None)
ax_idx += 1
fig.suptitle(title)
fig.set_size_inches(16, 16)
fig.tight_layout()
return fig
def plot_press_chronic():
global cached_dfs
de = None
if 'DE-total' in cached_dfs:
de = cached_dfs['DE-total']
else:
de = entorb.to_dataframe('DE-total')
rs1 = polynomial_r(de)
rs2 = weekly_r(de)
news = pd.read_csv('data/chronic_de.tsv',
sep="\\t",
usecols=['Datum', 'Ereignis'],
engine='python')
news['Datum'] = pd.to_datetime(news['Datum'], format='%Y-%m-%d')
news = news.set_index('Datum')
fig, ax = plt.subplots()
rs1['Cases'].transpose().plot(ax=ax, label='Logistisch')
rs2['Cases'].transpose().plot(ax=ax,
secondary_x=True,
label='Wochenfälle je 100 000 Ew.')
ax.set_title("Fall-Rate. COVID-19-Chronik der FAZ")
ax.set_yticks(news.index)
ax.set_ylim(news.index.max() + timedelta(days=4), rs1['Cases'].index.min())
ax.set_yticklabels(news['Ereignis'])
ax.grid()
ax.set_ylabel('')
plt.legend(loc='lower right')
fig.set_size_inches(9, 16)
return fig
def plot_weekly_r(col='Cases', ncols=4):
"""
Plot of <col> per week
Args:
col: What to plot. 'Cases' by default.
ncols: Columns of charts of resulting figure.
Return:
Figure
"""
global cached_dfs
areas = sorted([x for x in DE_STATE_NAMES])
fig, axes = plt.subplots(nrows=4, ncols=4, sharex=True, sharey=True)
for i, (ax, area) in enumerate(zip(axes.flat, areas)):
de = None
if area in cached_dfs:
de = cached_dfs[area]
else:
de = entorb.to_dataframe(area)
cached_dfs[area] = de
rs = weekly_r(de, DE_STATE_POPULATION[DE_STATE_NAMES[area]])
rs[col].plot(ax=ax, title=DE_STATE_NAMES[area])
fig.suptitle("Weekly new cases")
fig.set_size_inches(16, 16)
fig.tight_layout()
return fig
def plot_r(col='Cases', population=DE_POPULATION):
global cached_dfs
lasts = []
lasts_rki = []
areas = sorted([x for x in DE_STATE_NAMES])
fig, axes = plt.subplots(nrows=4, ncols=4, sharex=True, sharey=True)
for i, (ax, area) in enumerate(zip(axes.flat, areas)):
de = None
if area in cached_dfs:
de = cached_dfs[area]
else:
de = entorb.to_dataframe(area).rolling('7D').mean()
cached_dfs[area] = de
rs = polynomial_r(de, population[DE_STATE_NAMES[area]])
lasts.append(rs[col].tail(1).values[0])
ax = rs[col].plot(
ax=ax,
title="%s (%d Ew.)" %
(DE_STATE_NAMES[area], population[DE_STATE_NAMES[area]]))
rs = rki_r(de)
lasts_rki.append(rs[col].tail(1).values[0])
rs[col].plot(ax=ax)
fig.set_size_inches(16, 16)
fig.tight_layout()
return fig, lasts, lasts_rki
def plot_source_deltas():
fig, axes = plt.subplots(nrows=6)
da = entorb.to_dataframe(nation='DE')
db = risklayer.to_dataframe(nation='DE')
for ax, col in zip(axes.flat[:2], ['Cases', 'Cases_New']):
(da[col] - db[col]).plot(kind='bar',
label="%s, delta entorb - risklayer" % col,
sharex=True,
ax=ax)
ax.legend()
dc = rki.to_dataframe(nation='DE')
for ax, col in zip(axes.flat[2:4], ['Cases', 'Cases_New']):
(da[col] - dc[col]).plot(kind='bar',
label="%s, delta entorb - rki" % col,
sharex=True,
ax=ax)
ax.legend()
for src in [entorb, rki, risklayer]:
df = {entorb: da, rki: dc, risklayer: db}[src]
for ax, col in zip(axes.flat[4:], ['Cases', 'Cases_New']):
df[col].plot(label="%s, %s" % (col, src.__name__),
sharex=True,
ax=ax)
ax.legend()
fig.set_size_inches(16, 9)
return fig
def main():
import sys
if len(sys.argv) < 3:
print("USAGE: %s OUTFILEprojection1 OUTFILEfuture_unld2" % sys.argv[0])
sys.exit(2)
de = entorb.to_dataframe('DE-total')
with plt.style.context('ggplot'):
plot_projection(de, DE_POPULATION).savefig(sys.argv[1],
bbox_inches='tight')
DE_STATE_NAMES,
plot_projections([n for n in DE_STATE_NAMES], \
[entorb.to_dataframe(s) for s in DE_STATE_NAMES], \
[DE_STATE_POPULATION[DE_STATE_NAMES[s]] for s in DE_STATE_NAMES]) \
.savefig(sys.argv[2], bbox_inches='tight')
plot_projection(de=entorb.to_dataframe(nation='US'),
population=329e6).savefig(sys.argv[3],
bbox_inches='tight')
def plot_correlation():
# https://query.wikidata.org/
dictlist = return_sparql_query_results("""
SELECT DISTINCT ?code ?population ?area ?ppp ?ngdp ?growth ?totrsv ?hdi ?medinc ?literacy ?life_expectancy ?fertility_rate
{
?country wdt:P31 wd:Q3624078 ;
wdt:P297 ?code ;
wdt:P2046 ?area ;
wdt:P1082 ?population .
OPTIONAL {
?country wdt:P4010 ?ppp .
?country wdt:P2131 ?ngdp .
?country wdt:P2219 ?growth .
?country wdt:P2134 ?totrsv .
?country wdt:P1081 ?hdi .
?country wdt:P3529 ?medinc .
?country wdt:P6897 ?literacy .
?country wdt:P2250 ?life_expectancy .
?country wdt:P4841 ?fertility_rate .
}
}
""")['results']['bindings']
df = pd.DataFrame({
k: [x[k]['value'] if k is 'code' else float(x[k]['value']) if k in x else None for x in dictlist] \
for k in ['code', 'population', 'area', 'ppp', 'ngdp', 'growth', 'totrsv', 'hdi', 'medinc', 'literacy', 'life_expectancy', 'fertility_rate']}) \
.set_index('code')
df['density'] = df['population'] / df['area']
df['ngdp/p'] = df['ngdp'] / df['population']
df['ppp/p'] = df['ppp'] / df['population']
df['growth/p'] = df['growth'] / df['population']
df['totrsv/p'] = df['totrsv'] / df['population']
df['hdi/p'] = df['hdi'] / df['population']
cols = ['Deaths_Per_Million']
dict_entorb = {c: [] for c in cols}
for area in df.index:
try:
for col in cols:
dict_entorb[col].append(
entorb.to_dataframe(nation=area)[col].values[-1])
except:
for col in cols:
dict_entorb[col].append(None)
for col in dict_entorb:
df[col] = dict_entorb[col]
ncols = int((len(df.columns) - 1) / 4 + 1)
fig, axes = plt.subplots(ncols=ncols, nrows=4)
for ax, col in zip(axes.flat, [c for c in df.columns if c not in cols]):
df.plot(kind='scatter',
y='Deaths_Per_Million',
logy=True,
x=col,
sharey=False,
ax=ax)
fig.set_size_inches(16, 4 * ncols)
return fig
def plot_rki_and_logistic_total(state='DE-total'):
"""
Plot rki and logistic rates for Germany.
Args:
state: ISO code of German district, DE-total or Country
Return:
Figure
"""
global cached_dfs
de = None
if state not in cached_dfs:
de = entorb.to_dataframe(state)
cached_dfs[state] = de
else:
de = cached_dfs[state]
fig, ax = plt.subplots()
ax2 = ax.twinx()
l = ax2.plot(de.index,
de['Cases_New'].rolling('7D').mean(),
color='k',
label='New Cases rolling week')
ax2.plot(de.index,
de['Cases_New'],
linestyle=':',
color=l[0].get_color(),
label='New Cases')
poly = polynomial_r(de)['Cases']
l = ax.plot(de.index,
poly.rolling('7D').mean(),
label='Logistic rate rolling week')
ax.plot(de.index,
poly,
linestyle=':',
color=l[0].get_color(),
label='Logistic rate')
rki = rki_r(de)['Cases']
l = ax.plot(de.index,
rki.rolling('7D').mean(),
label='RKI rate rolling week')
ax.plot(de.index,
rki,
linestyle=':',
color=l[0].get_color(),
label='RKI rate')
plt.legend(handles=ax.lines + ax2.lines)
ax.axhline(1.0, color='g', alpha=0.5)
fig.set_size_inches(16, 9)
fig.tight_layout()
return fig
def plot_rki_and_logistic(col='Cases',
ncols=4,
population=DE_STATE_POPULATION):
"""
Plot of reproduction of <col>
Args:
col: What to plot. 'Cases' by default.
ncols: Columns of charts of resulting figure.
Return:
Tuple of a Figure, a list of last logistical rates and a list of last rki values
"""
global cached_dfs
areas = sorted([x for x in DE_STATE_NAMES])
lasts = {'area': areas, 'logistic': [], 'rki': [], 'weekly': []}
fig, axes = plt.subplots(nrows=4, ncols=4, sharex=True, sharey=True)
for i, (ax, area) in enumerate(zip(axes.flat, areas)):
de = None
if area in cached_dfs:
de = cached_dfs[area].rolling('7D').mean()
else:
cached_dfs[area] = entorb.to_dataframe(area)
de = cached_dfs[area].rolling('7D').mean()
rs = polynomial_r(de, population[DE_STATE_NAMES[area]])
lasts['logistic'].append(rs[col].values[-1])
rs[col].plot(ax=ax,
label="logistic",
title="%s (%d Ew." %
(DE_STATE_NAMES[area], population[DE_STATE_NAMES[area]]))
rs = rki_r(de)
lasts['rki'].append(rs[col].values[-1])
rs[col].plot(ax=ax, ylim=(1, 2), label="rki", sharex=True, sharey=True)
rs = weekly_r(de, population[DE_STATE_NAMES[area]])
lasts['weekly'].append(rs[col].values[-1])
#ax2 = rs[col].plot(ax=ax, label="weekly", sharex=True, sharey=True, logy=True)
fig.set_size_inches(16, 16)
return fig, pd.DataFrame(lasts).set_index('area')
def plot_history(dataframe, fname):
"""Plot each federal state's total exhaustion from @entorb but also keep each curve in
dict bundesland_curves for later correlations"""
availability = dict()
deaths = dict()
df = dataframe.set_index('Stand').bfill()
for federal_state in federal_state_translation:
availability[federal_state] = df[df['Bundesland'] ==
federal_state].resample('D').mean()
deaths[federal_state] = entorb.to_dataframe(
federal_state)['Deaths_New']
deaths[federal_state].name = 'Deaths'
for federal_state in sorted(list(federal_state_translation.keys())):
fig, ax = plt.subplots()
ax = deaths[federal_state].plot(kind='line',
ax=ax,
color='grey',
linestyle=':',
marker='+',
legend=True,
figsize=(16, 9))
ax.set_xlabel('Deaths')
ax = availability[federal_state].plot(
kind='line', ax=ax, secondary_y=True,
marker='') # drawstyle=steps-post
ax.set_yticks([
0.0, # * num_clinics[federal_state],
0.5, # * num_clinics[federal_state],
1.0, # * num_clinics[federal_state],
])
ax.set_yticklabels(["available", "limited", "unavailable"])
ax.set_title("%s" % federal_state_translation[federal_state])
fig.savefig(fname % {'state': federal_state})
def main():
import sys
area = 'DE-total'
de = entorb.to_dataframe(area)
