-
Notifications
You must be signed in to change notification settings - Fork 1
/
forecast.py
545 lines (398 loc) · 18.4 KB
/
forecast.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Mon Dec 20 19:21:46 2021
@author: eduardoaraujo
"""
# importando os pacotes necessários
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
import io
from datetime import timedelta
from io import BytesIO
import base64
import json
import pickle
import uuid
import re
import plotly.graph_objects as go
import streamlit as st
from get_data import get_canton_data, get_curve
from plots import scatter_plot_cases_hosp
from sqlalchemy import create_engine
engine = create_engine(
"postgresql://epigraph:epigraph@localhost:5432/epigraphhub")
def plot_predictions(table_name, curve, title=None):
''''
Function to plot the predictions
params table_name: Name of the table with the predictions (name used to save
the table in the database)
params curve: Curve related with the predictions that will be plotted
return plotly figure
'''
target_curve_name = curve
canton = 'GE'
df_val = pd.read_sql_table(
table_name, engine, schema='switzerland', index_col='date')
df_val = df_val.loc[df_val.canton == canton]
target = df_val['target']
train_size = df_val['train_size'].values[0]
x = df_val.index
y5 = df_val['lower']
y50 = df_val['median']
y95 = df_val['upper']
point = target.index[train_size]
min_val = min([min(target), np.nanmin(y95)])
max_val = max([max(target), np.nanmax(y95)])
point_date = np.where(target.index == '2021-01-01')
fig = go.Figure()
# Dict with names for the curves
names = {'hosp': 'New Hospitalizations', 'icu_patients': 'Total ICU patients',
'total_hosp': 'Total hospitalizations'}
if title == None:
title = f"{canton}"
fig.update_layout(width=900, height=500, title={
'text': title,
'y': 0.87,
'x': 0.42,
'xanchor': 'center',
'yanchor': 'top'},
xaxis_title='Date',
yaxis_title=f'{names[target_curve_name]}',
template='plotly_white')
# adding the traces
# Data
fig.add_trace(go.Scatter(x=target.index, y=target.values,
name='Data', line=dict(color='black')))
# Line separing training data and test data
fig.add_trace(go.Scatter(x=[point, point], y=[
min_val, max_val], name="Out of sample predictions", mode='lines', line=dict(color='#1CA71C', dash='dash')))
# Separação entre os dados de teste e o forecast
# fig.add_trace(go.Scatter(x=[target.index[-1], target.index[-1]], y=[min_val,max_val], name="Forecast", mode = 'lines',line=dict(color = '#FB0D0D', dash = 'dash')))
# NGBoost predictions
fig.add_trace(go.Scatter(x=x, y=y50, name='NGBoost',
line=dict(color='#FF7F0E')))
fig.add_trace(go.Scatter(x=x, y=y5, line=dict(
color='#FF7F0E', width=0), showlegend=False))
fig.add_trace(go.Scatter(x=x, y=y95, line=dict(color='#FF7F0E', width=0),
mode='lines',
fillcolor='rgba(255, 127, 14, 0.3)', fill='tonexty', showlegend=False))
fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
return fig
def plot_cases():
''''
Function to plot the new cases according to FOPH in GE
return[0] plotly figure
return[1] last data of new cases reported
'''
df = get_curve('cases', 'GE')
last_date = df.index[-1]
df = df['2021-11-01':]
#df = df.iloc[:-3]
# computing the rolling average
m_movel = df.rolling(7).mean().dropna()
fig = go.Figure()
title = "Geneva"
fig.update_layout(width=900, height=500, title={
'text': title,
'y': 0.87,
'x': 0.42,
'xanchor': 'center',
'yanchor': 'top'},
legend={'orientation': 'h', 'valign':'top', 'y': -0.25},
xaxis_title='Report Date',
yaxis_title='New cases',
template='plotly_white')
fig.add_trace(go.Bar(x=df.index, y=df.entries, name='New cases',
marker_color='rgba(31, 119, 180, 0.7)'))
fig.add_trace(go.Scatter(x=m_movel.index, y=m_movel.entries,
name='Rolling average', line=dict(color='black', width=2)))
fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
return fig, last_date, df.entries[-2:]
def get_hospCapacity():
df = get_curve('hospcapacity', 'GE')
df = df.resample('D').mean()
df = df.sort_index()
#df = df.iloc[:-3]
return df.total_covid19patients[-2:].astype('int'),df.totalpercent_covid19patients[-2:]
def plot_hosp():
''''
Function to plot the number of new hospitalizations for GE
returns plotly figure
'''
df = get_curve('hosp', 'GE')
df = df['2021-11-01':]
#df = df.iloc[:-3]
# computing the rolling average
m_movel = df.rolling(7).mean().dropna()
fig = go.Figure()
title = "Geneva"
fig.update_layout(width=900, height=500, title={
'text': title,
'y': 0.87,
'x': 0.42,
'xanchor': 'center',
'yanchor': 'top'},
legend={'orientation': 'h', 'valign':'top', 'y': -0.25},
xaxis_title='Report Date',
yaxis_title='New hospitalizations',
template='plotly_white')
fig.add_trace(go.Bar(x=df.index, y=df.entries,
name='New hospitalizations', marker_color='rgba(31, 119, 180, 0.7)'))
fig.add_trace(go.Scatter(x=m_movel.index, y=m_movel.entries,
name='Rolling average', line=dict(color='black', width=2)))
fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
return fig, df.entries[-2:]
def plot_forecast(table_name, curve, title=None):
''''
Function to plot the forecast
params table_name: Name of the table with the predictions (name used to save
the table in the database)
params curve: Curve related with the predictions that will be plotted
return[0] plotly figure
return[1] dataframe with the values forecasted
'''
target_curve_name = curve
canton = 'GE'
df_for = pd.read_sql_table(
table_name, engine, schema='switzerland', index_col='date')
df_for = df_for.loc[df_for.canton == canton]
#if table_name == 'ml_forecast_hosp_up':
#ydata = get_updated_data(smooth=True)
curves = {'hosp': 'hosp', 'icu_patients': 'hospcapacity', 'total_hosp': 'hospcapacity'}
ydata = get_canton_data(curves[curve], ['GE'])
ydata = ydata.resample('D').mean()
#ydata = ydata.iloc[:-3]
ydata = ydata.rolling(7).mean().dropna()
dates_forecast = df_for.index
forecast5 = df_for['lower']
forecast50 = df_for['median']
forecast95 = df_for['upper']
fig = go.Figure()
# Dict with names for the curves
names = {'hosp': 'Forecast New Hospitalizations',
'icu_patients': 'Forecast Total ICU patients',
'total_hosp': 'Total hospitalizations'}
if title == None:
title = f"{canton}"
fig.update_layout(width=900, height=500, title={
'text': title,
'y': 0.87,
'x': 0.42,
'xanchor': 'center',
'yanchor': 'top'},
xaxis_title='Date',
yaxis_title=f'{names[target_curve_name]}',
template='plotly_white')
column_curves = {'hosp': 'entries',
'icu_patients': 'icu_covid19patients',
'total_hosp': 'total_covid19patients'}
min_data = min(ydata.index[-1], df_for.index[0] - timedelta(days=1))
fig.add_trace(go.Scatter(
x=ydata.loc[:min_data].index[-150:], y=ydata.loc[:min_data][column_curves[curve]][-150:], name='Data', line=dict(color='black')))
# Separation between data and forecast
fig.add_trace(go.Scatter(x=[df_for.index[0], df_for.index[0]], y=[min(min(ydata[column_curves[curve]][-150:]), min(forecast95)), max(
max(ydata[column_curves[curve]][-150:]), max(forecast95))], name="Data/Forecast", mode='lines', line=dict(color='#FB0D0D', dash='dash')))
# NGBoost
fig.add_trace(go.Scatter(x=dates_forecast, y=forecast50,
name='Forecast NGBoost', line=dict(color='#FF7F0E')))
fig.add_trace(go.Scatter(x=dates_forecast, y=forecast5, line=dict(
color='#FF7F0E', width=0), mode='lines', showlegend=False))
fig.add_trace(go.Scatter(x=dates_forecast, y=forecast95, line=dict(color='#FF7F0E', width=0),
mode='lines',
fillcolor='rgba(255, 127, 14, 0.3)', fill='tonexty', showlegend=False))
fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgray', zeroline=False,
showline=True, linewidth=1, linecolor='black', mirror=True)
# if path == None:
#path = f"images/forecast_{canton}.png"
# fig.write_image(path)
#del df_for['index']
df_for.index = pd.to_datetime(df_for.index)
df_for.index = df_for.index.date
df_for.reset_index(inplace=True)
df_for.rename(columns={'index': 'date'}, inplace=True)
return fig, df_for
def download_button(object_to_download, download_filename, button_text, pickle_it=False):
"""
Generates a link to download the given object_to_download.
Params:
------
object_to_download: The object to be downloaded.
download_filename (str): filename and extension of file. e.g. mydata.csv,
some_txt_output.txt download_link_text (str): Text to display for download
link.
button_text (str): Text to display on download button (e.g. 'click here to download file')
pickle_it (bool): If True, pickle file.
Returns:
-------
(str): the anchor tag to download object_to_download
Examples:
--------
download_link(your_df, 'YOUR_DF.csv', 'Click to download data!')
download_link(your_str, 'YOUR_STRING.txt', 'Click to download text!')
"""
if pickle_it:
try:
object_to_download = pickle.dumps(object_to_download)
except pickle.PicklingError as e:
st.write(e)
return None
else:
if isinstance(object_to_download, bytes):
pass
elif isinstance(object_to_download, pd.DataFrame):
#object_to_download = object_to_download.to_csv(index=False)
towrite = io.BytesIO()
object_to_download = object_to_download.to_excel(
towrite, encoding='utf-8', index=False, header=True)
towrite.seek(0)
# Try JSON encode for everything else
else:
object_to_download = json.dumps(object_to_download)
try:
# some strings <-> bytes conversions necessary here
b64 = base64.b64encode(object_to_download.encode()).decode()
except AttributeError as e:
b64 = base64.b64encode(towrite.read()).decode()
button_uuid = str(uuid.uuid4()).replace('-', '')
button_id = re.sub('\d+', '', button_uuid)
custom_css = f"""
<style>
#{button_id} {{
display: inline-flex;
align-items: center;
justify-content: center;
background-color: rgb(255, 255, 255);
color: rgb(38, 39, 48);
padding: .25rem .75rem;
position: relative;
text-decoration: none;
border-radius: 4px;
border-width: 1px;
border-style: solid;
border-color: rgb(230, 234, 241);
border-image: initial;
}}
#{button_id}:hover {{
border-color: rgb(246, 51, 102);
color: rgb(246, 51, 102);
}}
#{button_id}:active {{
box-shadow: none;
background-color: rgb(246, 51, 102);
color: white;
}}
</style> """
dl_link = custom_css + \
f'<a download="{download_filename}" id="{button_id}" href="data:application/vnd.openxmlformats-officedocument.spreadsheetml.sheet;base64,{b64}">{button_text}</a><br></br>'
return dl_link
def app():
st.title('Switzerland COVID-19 Hospitalization forecasts')
fig_c, last_date, last_cases = plot_cases()
fig_h, last_hosp = plot_hosp()
total_hc, perc_hc = get_hospCapacity()
st.markdown(f'''
## Current Status in Geneva
On **{last_date.date()}**, the FOPH (Federal Office of Public Health) reported:
''')
c1, c2 = st.columns(2)
with c1:
st.metric("Daily new cases", value=last_cases[-1],
delta=f"{last_cases.diff()[-1]} cases", delta_color="inverse")
st.metric("Total COVID-19 Hospitalizations", value=total_hc[-1],
delta=f"{total_hc.diff()[-1]} cases", delta_color='inverse')
st.plotly_chart(fig_c, use_container_width=True)
# print(last_cases,last_cases.diff())
with c2:
st.metric("Daily new Hospitalizations", value=last_hosp[-1],
delta=f"{last_hosp.diff()[-1]:.1f} Hospitalizations", delta_color="inverse")
st.metric("Percent of total Hospitalizations", value=f"{perc_hc[-1]:.2f}",
delta=f"{perc_hc.diff()[-1]:.2f} %",
delta_color="inverse")
st.plotly_chart(fig_h, use_container_width=True)
st.write('''
For forecasts of other cantons, see sidebar menu.
''')
st.write('''
#### Relation between cases and hospitalizations in Geneva:
If we look at hospitalization *vs* cases, we can hint at the change in case severity over time.
''')
scatter_cases_hosp_GE = scatter_plot_cases_hosp('GE')
st.image(scatter_cases_hosp_GE)
st.title('Forecasts')
st.write('''
To forecast the daily hospitalizations, total hospitalizations, and total ICU hospitalizations
in canton Geneva, it was used a Gradient Boosting Machine that returns
probabilistic predictions implemented in the python package `NGBoost`.
In the model, we use as predictors the series of cases, hospitalizations,
tests and ICU occupations from all the cantons belonging to the same cluster
of the one we are forecasting for, as well as total vaccinations per hundred thousand in
Switzerland. The regression, in somewhat compact notation, is defined as ''')
st.latex(r'''
H_{k,t} \sim C_{k,t-\tau_i} + H_{k,t-\tau_i} +V_{k,t-\tau_i} + ICU_{k,t-\tau_i},
''')
st.write(r'''
for the daily hospitalizations, where C stands for cases, H stands for hospitalizations, V for vaccination,
and ICU for the number of ICU patients. The model bases its predictions
on the last 14 days: $\tau_1, \tau_2, \ldots, \tau_{14}$ and predicts the
next 14 days. For each of these 14 days, one model is trained. The regression for the
total hospitalizations and total ICU hospitalizations follows the same format defined above.
''')
st.write('''
## 14-day Forecasts
Below, we have the forecast for the next 14 days, for daily hospitalizations, total hospitalizations
and total ICU hospitalizations. The 95% confidence bounds are also shown. The table with the forecasts can be downloaded by clicking on the button.
''')
st.write('## Forecast results')
st.write('### New Hospitalizations')
fig_for, df_hosp = plot_forecast('ngboost_forecast_hosp_d_results', curve='hosp')
st.plotly_chart(fig_for, use_container_width=True)
filename = 'forecast_hosp.csv'
download_button_str = download_button(
df_hosp, filename, 'Download data', pickle_it=False)
st.markdown(download_button_str, unsafe_allow_html=True)
st.write('### Total Hospitalizations')
fig_for, df_total = plot_forecast('ngboost_forecast_total_hosp_d_results', curve='total_hosp')
st.plotly_chart(fig_for, use_container_width=True)
filename = 'forecast_total_hosp.csv'
download_button_str = download_button(
df_total, filename, 'Download data', pickle_it=False)
st.markdown(download_button_str, unsafe_allow_html=True)
st.write('### Total ICU Hospitalizations')
fig_for, df_icu = plot_forecast('ngboost_forecast_icu_patients_d_results', curve='icu_patients')
st.plotly_chart(fig_for, use_container_width=True)
filename = 'forecast_ICU.csv'
download_button_str = download_button(
df_icu, filename, 'Download data', pickle_it=False)
st.markdown(download_button_str, unsafe_allow_html=True)
st.write('''
## Model Validation
In the Figure below, the model's predictions are plotted against data, both *in sample* (for
the data range used for training) and *out of sample* (part of the series not used during model training)
for the machine learning model.
''')
fig = plot_predictions('ngboost_validation_hosp_d_results', curve='hosp')
st.plotly_chart(fig, use_container_width=True)
st.write('''
Below, we have the same as above, but for the total hospitalizations.
''')
fig = plot_predictions('ngboost_validation_total_hosp_d_results', curve='total_hosp')
st.plotly_chart(fig, use_container_width=True)
st.write('''
Below, we have the same as above, but for the ICU occupancy.
''')
fig = plot_predictions('ngboost_validation_icu_patients_d_results', curve='icu_patients')
st.plotly_chart(fig, use_container_width=True)