hospital_data.py

   1 #!/usr/bin/env python
   2 # coding: utf-8
   3 # %%
   4 import itertools
   5 import collections
   6 import json
   7 import pandas as pd
   8 import numpy as np
   9 from scipy.stats import gmean
  10 import datetime
  11
  12 import sqlalchemy
  13
  14 import matplotlib as mpl
  15 import matplotlib.pyplot as plt
  16 plt.ioff()
  17
  18
  19 # %%
  20 connection_string = 'postgresql://covid:3NbjJTkT63@localhost/covid'
  21
  22
  23 # %%
  24 engine = sqlalchemy.create_engine(connection_string)
  25
  26
  27 # %%
  28 normalisation_date = '2020-08-01' # '2020-05-15'
  29
  30 hospital_normalisation_date = {
  31     'hospital_normalisation_date': normalisation_date
  32 }
  33
  34 with open('hospital_normalisation_date.json', 'w') as f:
  35     json.dump(hospital_normalisation_date, f)
  36
  37 hnd = pd.to_datetime(normalisation_date).strftime("%d %B %Y")
  38 with open('hospital_normalisation_date.js', 'w') as f:
  39     f.write(f"document.write('{hnd}');")
  40
  41
  42 # %%
  43 qstr = '''select uk_data.date
  44     , uk_data.new_admissions, uk_data_7.new_admissions as new_admissions_7
  45     , uk_data.new_cases, uk_data_7.new_cases as new_cases_7
  46     , uk_data.new_deaths, uk_data_7.new_deaths as new_deaths_7
  47     , uk_data.hospital_cases, uk_data_7.hospital_cases as hospital_cases_7
  48     , uk_data.ventilator_beds, uk_data_7.ventilator_beds as ventilator_beds_7
  49     from uk_data left outer join uk_data_7 using (date)
  50     order by uk_data.date'''
  51 hospital_data = pd.read_sql_query(qstr, engine,
  52     index_col='date',
  53     parse_dates = ['date'])
  54
  55
  56 # %%
  57 fig = plt.figure(figsize=(10, 8))
  58 ax = hospital_data.loc['2020-03-20':, ['hospital_cases', 'ventilator_beds']].plot(
  59     secondary_y='ventilator_beds',
  60     figsize=(10, 8),
  61     title="People in hospital, and on mechanical ventilation")
  62 # axl.legend(['Number in hospital', 'Number on ventilator'])
  63 plt.savefig('people_in_hospital.png')
  64
  65
  66 # %%
  67 hospital_data['cases_m7nd'] = hospital_data.new_cases_7 / hospital_data.new_cases_7.loc[normalisation_date]
  68 hospital_data['deaths_m7nd'] = hospital_data.new_deaths_7 / hospital_data.new_deaths_7.loc[normalisation_date]
  69 hospital_data['admissions_m7nd'] = hospital_data.new_admissions_7 / hospital_data.new_admissions_7.loc[normalisation_date]
  70
  71
  72 # %%
  73 ymax = (int(hospital_data.loc[normalisation_date:, ['cases_m7nd', 'admissions_m7nd', 'deaths_m7nd']].max().max() * 10) + 1) / 10.0
  74 ax = hospital_data.loc[normalisation_date:, ['cases_m7nd', 'admissions_m7nd', 'deaths_m7nd']].plot(
  75     ylim=(0, ymax),
  76     figsize=(10, 8),
  77     title="Cases, hospital admissions, and deaths\n(7-day moving averages)")
  78
  79 lvi = hospital_data.cases_m7nd.last_valid_index()
  80 ax.text(x = lvi + pd.Timedelta(days=1), y = hospital_data.cases_m7nd[lvi], s = f'{hospital_data.cases_m7nd[lvi]:.2f}')
  81
  82 lvi = hospital_data.admissions_m7nd.last_valid_index()
  83 ax.text(x = lvi + pd.Timedelta(days=1), y = hospital_data.admissions_m7nd[lvi], s = f'{hospital_data.admissions_m7nd[lvi]:.2f}')
  84
  85 lvi = hospital_data.deaths_m7nd.last_valid_index()
  86 ax.text(x = lvi + pd.Timedelta(days=1), y = hospital_data.deaths_m7nd[lvi], s = f'{hospital_data.deaths_m7nd[lvi]:.2f}')
  87
  88
  89 ax.set_xlabel("Date")
  90 ax.set_ylabel(f'Normalised units, scaled from {normalisation_date}')
  91 ax.legend(['Cases', 'Admissions', 'Deaths'])
  92
  93 plt.savefig('cases_admissions_deaths.png')
  94
  95
  96 # %%
  97
  98
  99
 100