MJacobs1985

rm(list = ls())
library(tidyverse)
library(readr)
library(ggthemes)
library(lubridate)
library(zoo)
library(gganimate)
getwd()
data_folder <- file.path("C:/Users/marcj/Nutreco/StatisticsPlatform/Website/ELearning/Workshops/CausalAnalysis/Covid/")
url <- "https://covid.ourworldindata.org/data/owid-covid-data.csv"

MJacobs1985

rm(list = ls())
library(tidyverse)
library(readr)
library(ggthemes)
library(lubridate)
library(zoo)
getwd()
data_folder <- file.path("")
url <- "https://covid.ourworldindata.org/data/owid-covid-data.csv"
name <- "owid-covid-data.csv"

MJacobs1985

library(readxl)
library(ggplot2)
library(viridis)
library(tidyverse)
UK <- read_excel("referencetable.xlsx", sheet = "Table 7", skip = 4, n_max = 350)
UK$Mortality.age.std=as.numeric(UK$Mortality.age.std)
UK$Lower.limit=as.numeric(UK$Lower.limit)
UK$Upper.limit=as.numeric(UK$Upper.limit)
mynamestheme=theme(plot.title=element_text(face="bold",size=(15)))
UK.covid <- read_excel("referencetable.xlsx", sheet = "Table 5", skip = 4, n_max = 350)

MJacobs1985

proc sort data=try2 out=Work.preProcessedData;by date;run;
proc varmax data=Work.preProcessedData plots(only)=(forecasts impulse model residual) 
outest=work.outest covout outstat=work.outstat;
id date interval=day;
where '22MAR2020'd <= date <= '9DEC2021'd;
model 
grocery_and_pharmacy 
parks 
residential 
retail_and_recreation 

MJacobs1985

/* LETS LOOK AT THE WORLD DATA */ 
FILENAME REFFILE '/Data/owid-covid-data.xlsx';
PROC IMPORT DATAFILE=REFFILE DBMS=XLSX OUT=WORK.IMPORT;GETNAMES=YES;RUN;
proc freq data=import nlevels;table date;run;
proc freq data=import nlevels;table location;run;
data import; set import; 
new = input(date,yymmdd10.);
drop date;
rename new=date;
run;

MJacobs1985

options(mc.cores=parallel::detectCores())
model4<- stan_glmer(
  ADG~ADF + 
    Challenge + 
    Included2017 + 
    BW_start+Country+
    n_start+
    (ADF|Studycode),
  data = d, family = gaussian,

MJacobs1985

predictions_complete <- posterior_predict(model3,newdata = d)
ppc_intervals(d$ADG, 
              yrep = predictions_complete,
              prob_outer = 0.80) +
  ggplot2::facet_grid(~d$Studycode, scales="free")+
  xaxis_text(angle = 90, hjust = 1)

MJacobs1985

Book<- read_excel("X:/StatisticsPlatform/Website/ELearning/Workshops/Bayesian Analysis/Data/Book1.xlsx")
d<-Book%>%dplyr::select(ADG,ADF,Country,Days,BW_start,n_start,Challenge,Studycode,Included2017)
d<-d[complete.cases(d), ]
d$Studycode<-as.factor(d$Studycode)
d$Challenge<-as.factor(d$Challenge)
d$Country<-as.factor(d$Country)

model3<- stan_glmer(
  ADG~ADF + 

MJacobs1985

posterior_interval(model2, prob = 0.80)
d %>% 
  add_fitted_draws(model2, n = 20) %>%
  ggplot(aes(x = ADF, y = ADG)) +
  geom_line(aes(y = .value, group = paste(Studycode, .draw))) + 
  facet_wrap(~ .draw) +theme_bw()
d %>%
  add_predicted_draws(model2, n = 100) %>%
  ggplot(aes(x = ADF)) +
  geom_density(aes(x = .prediction, group = .draw)) +

MJacobs1985

predictions <- posterior_predict(model2, newdata = d)
ppc_intervals(d$ADG, 
              yrep = predictions, 
              x = d$ADF, 
              prob = 0.5, prob_outer = 0.95)
prediction_summary(model2, data = d)
cv_procedure <- prediction_summary_cv(model = model2,
                                      group = "Studycode",
                                      data  = d, 
                                      k = 10)