Convert summarised quantiles from wide to long format
Arguments
- posterior
A dataframe as output by
fv_tidy_posterior(),fv_extract_forecast(), etc.
See also
Functions used for postprocessing of model fits
convert_to_stanfit(),
extract_draws(),
extract_forecast_dates(),
fv_extract_forecast(),
fv_posterior(),
fv_tidy_posterior(),
link_dates_with_posterior(),
link_obs_with_posterior(),
plot.fv_posterior(),
print.fv_posterior(),
summary.fv_posterior(),
update_voc_label()
Examples
posterior <- fv_example(strains = 2, type = "posterior")
long_posterior <- quantiles_to_long(posterior)
long_posterior
#> value_type variable clean_name date type obs
#> 1: model avg_voc_advantage Average VOC effect <NA> NA
#> 2: model beta Beta <NA> NA
#> 3: model init_cases[1] Initial cases <NA> NA
#> 4: model init_voc_cases[1] Initial VOC cases <NA> NA
#> 5: model phi[1] Notification overdispersion <NA> NA
#> ---
#> 1888: raw log_lik[11] <NA> NA
#> 1889: raw log_lik[12] <NA> NA
#> 1890: raw log_lik[13] <NA> NA
#> 1891: raw log_lik[14] <NA> NA
#> 1892: raw log_lik[15] <NA> NA
#> observed forecast_start exponentiated mean median sd
#> 1: NA NA FALSE 0.589 0.589 3.08e-02
#> 2: NA NA FALSE 0.230 0.260 4.08e-01
#> 3: NA NA FALSE 89800.000 89600.000 6.14e+03
#> 4: NA NA FALSE 222.000 221.000 2.27e+01
#> 5: NA NA FALSE 95.300 78.000 7.42e+01
#> ---
#> 1888: NA NA NA -14.700 -14.600 6.93e-01
#> 1889: NA NA NA -13.800 -13.800 5.79e-01
#> 1890: NA NA NA -13.500 -13.400 8.67e-01
#> 1891: NA NA NA -8.160 -8.070 5.98e-01
#> 1892: NA NA NA -7.730 -7.590 6.97e-01
#> mad rhat ess_bulk ess_tail quantile prediction
#> 1: 2.97e-02 1.00 1270 1250 0.05 0.541
#> 2: 4.53e-01 1.00 779 923 0.05 -0.466
#> 3: 5.82e+03 1.00 1360 1340 0.05 79900.000
#> 4: 2.16e+01 1.00 1610 1060 0.05 186.000
#> 5: 5.32e+01 1.00 746 1030 0.05 21.200
#> ---
#> 1888: 6.23e-01 1.00 1300 1500 0.95 -13.700
#> 1889: 5.44e-01 1.01 915 1160 0.95 -13.000
#> 1890: 7.77e-01 1.00 1220 1490 0.95 -12.300
#> 1891: 4.91e-01 1.00 1070 1250 0.95 -7.420
#> 1892: 5.32e-01 1.00 910 1320 0.95 -6.880