def fit_doubling_time():
"""
fit time series for doubling time
"""
global d_countries_timeseries
global d_selected_countries
# TODO!
# l = []
# l.append('South Korea')
# for country in l:
for country in tqdm(d_countries_timeseries.keys()):
# for country in d_selected_countries.keys():
# print(country)
# country_code = d_selected_countries[country]['Code']
l_country_data = d_countries_timeseries[country]
# pop_in_Mill = d_selected_countries[country]['Population'] / 1000000
# for fits of doubling time
data_t = []
data_cases = []
data_deaths = []
for i in range(len(l_country_data)):
entry = l_country_data[i]
# for fits of doubling time
data_t.append(entry['Days_Past'])
data_cases.append(entry['Cases'])
data_deaths.append(entry['Deaths'])
l_country_data[i] = entry
# fit the doubling time each day
data = list(zip(data_t, data_cases))
fit_series_res_cases = helper.series_of_fits(data,
fit_range=7,
max_days_past=28)
data = list(zip(data_t, data_deaths))
fit_series_res_deaths = helper.series_of_fits(data,
fit_range=7,
max_days_past=28)
for i in range(len(l_country_data)):
entry = l_country_data[i]
entry['Cases_Doubling_Time'] = ""
entry['Deaths_Doubling_Time'] = ""
this_DaysPast = entry['Days_Past']
if this_DaysPast in fit_series_res_cases:
entry['Cases_Doubling_Time'] = fit_series_res_cases[
this_DaysPast]
if this_DaysPast in fit_series_res_deaths:
entry['Deaths_Doubling_Time'] = fit_series_res_deaths[
this_DaysPast]
l_country_data[i] = entry
if args["sleep"]:
time.sleep(1)
d_countries_timeseries[country] = l_country_data
def fit_doubling_or_halftime() -> dict:
for code, l_time_series in d_states_data.items():
print(f'fitting doubling time for {code}')
# if code != 'DE-total': # TODO
# continue
# # fit cases data V2: based on CasesNew instead of Cases and interpreting T<0 -> halftime
dataCases = []
for i in range(1, len(l_time_series)): # TODO
# for i in range(10, 60):
# x= day , y = cases
dataCases.append((
l_time_series[i]['Days_Past'],
l_time_series[i]['Cases_Last_Week_Per_100000']
# l_time_series[i]['Cases_New_Per_Million']
# this set to very noisy results, so using Last_week data instead
))
fit_series_res = helper.series_of_fits(dataCases,
fit_range=14,
max_days_past=365,
mode='exp')
for i in range(0, len(l_time_series)):
this_Doubling_Time = ""
this_days_past = l_time_series[i]['Days_Past']
if this_days_past in fit_series_res:
this_Doubling_Time = fit_series_res[this_days_past]
l_time_series[i][
'Cases_Last_Week_Doubling_Time'] = this_Doubling_Time
# debugging
# print(l_time_series[i]['Days_Past'], this_Doubling_Time)
d_states_data[code] = l_time_series
return d_states_data
def read_csv_to_dict() -> dict:
"""
read and convert the source csv file, containing: federalstate,infections,deaths,date,newinfections,newdeaths
re-calc _New
add _Per_Million
add Fitted Doubling time
"""
global d_states_ref
# Preparations
d_states_data = {'BW': [], 'BY': [], 'BE': [], 'BB': [], 'HB': [], 'HH': [], 'HE': [], 'MV': [
], 'NI': [], 'NW': [], 'RP': [], 'SL': [], 'SN': [], 'ST': [], 'SH': [], 'TH': []}
# add German sum
d_states_data['DE-total'] = []
d_german_sums = {} # date -> 'infections', 'deaths', 'new infections', 'new deaths'
# data body
with open(download_file, mode='r', encoding='utf-8') as f:
csv_reader = csv.DictReader(f, delimiter=",")
for row in csv_reader:
d = {}
s = row['date']
l = s.split("-")
d['Date'] = helper.date_format(
int(l[0]), int(l[1]), int(l[2]))
d['Cases'] = int(row["infections"])
d['Deaths'] = int(row["deaths"])
if row["federalstate"] == 'Baden-Württemberg':
d_states_data['BW'].append(d)
elif row["federalstate"] == 'Bavaria':
d_states_data['BY'].append(d)
elif row["federalstate"] == 'Berlin':
d_states_data['BE'].append(d)
elif row["federalstate"] == 'Brandenburg':
d_states_data['BB'].append(d)
elif row["federalstate"] == 'Bremen':
d_states_data['HB'].append(d)
elif row["federalstate"] == 'Hamburg':
d_states_data['HH'].append(d)
elif row["federalstate"] == 'Hesse':
d_states_data['HE'].append(d)
elif row["federalstate"] == 'Lower Saxony':
d_states_data['NI'].append(d)
elif row["federalstate"] == 'North Rhine-Westphalia':
d_states_data['NW'].append(d)
elif row["federalstate"] == 'Mecklenburg-Western Pomerania':
d_states_data['MV'].append(d)
elif row["federalstate"] == 'Rhineland-Palatinate':
d_states_data['RP'].append(d)
elif row["federalstate"] == 'Saarland':
d_states_data['SL'].append(d)
elif row["federalstate"] == 'Saxony':
d_states_data['SN'].append(d)
elif row["federalstate"] == 'Saxony-Anhalt':
d_states_data['ST'].append(d)
elif row["federalstate"] == 'Schleswig-Holstein':
d_states_data['SH'].append(d)
elif row["federalstate"] == 'Thuringia':
d_states_data['TH'].append(d)
else:
print("ERROR: unknown state")
quit()
# add to German sum
if d['Date'] not in d_german_sums:
d2 = {}
d2['Cases'] = d['Cases']
d2['Deaths'] = d['Deaths']
else:
d2 = d_german_sums[d['Date']]
d2['Cases'] += d['Cases']
d2['Deaths'] += d['Deaths']
d_german_sums[d['Date']] = d2
del d2
# for German sum -> same dict
for datum in d_german_sums.keys():
d = d_german_sums[datum]
d['Date'] = datum # add date field
d_states_data['DE-total'].append(d)
del d_german_sums, d
# check if DE sum of lastdate and per-last date has changed, if so: remove last date
if d_states_data['DE-total'][-1]['Cases'] == d_states_data['DE-total'][-2]['Cases']:
print("WARNING: DE cases sum is unchanged")
for code in d_states_data:
d_states_data[code].pop()
print(f"DE-States Last Date: {d_states_data['DE-total'][-1]['Date']}")
for code in d_states_data.keys():
l_time_series = d_states_data[code]
l_time_series = helper.prepare_time_series(l_time_series)
# add days past and per million
for i in range(len(l_time_series)):
d = l_time_series[i]
# add per Million rows
d = helper.add_per_million_via_lookup(d, d_states_ref, code)
# fit cases data
# Cases
dataCases = []
dataDeaths = []
for i in range(1, len(l_time_series)):
# x= day , y = cases
dataCases.append(
(
l_time_series[i]['Days_Past'],
l_time_series[i]['Cases']
)
)
dataDeaths.append(
(
l_time_series[i]['Days_Past'],
l_time_series[i]['Deaths']
)
)
fit_series_res = helper.series_of_fits(
dataCases, fit_range=7, max_days_past=28)
for i in range(0, len(l_time_series)):
this_Doubling_Time = ""
this_days_past = l_time_series[i]['Days_Past']
if this_days_past in fit_series_res:
this_Doubling_Time = fit_series_res[this_days_past]
l_time_series[i]['Cases_Doubling_Time'] = this_Doubling_Time
fit_series_res = helper.series_of_fits(
dataDeaths, fit_range=7, max_days_past=28)
for i in range(0, len(l_time_series)):
this_Doubling_Time = ""
this_days_past = l_time_series[i]['Days_Past']
if this_days_past in fit_series_res:
this_Doubling_Time = fit_series_res[this_days_past]
l_time_series[i]['Deaths_Doubling_Time'] = this_Doubling_Time
d_states_data[code] = l_time_series
if args["sleep"]:
time.sleep(1)
return d_states_data