push tests and data

.gitignore

@@ -0,0 +1,4 @@
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata

Example_logistic_reg_agg_data.R

@@ -0,0 +1,18 @@
+library(tidyverse)
+set.seed(5)
+df<-tibble(Gender = as.factor(sample(c("m","f"), 200, replace = TRUE, prob=c(0.6,0.4))),
+           Age_Group = as.factor(sample(c("[<30]","[30-65]", "[65+]"), 200, replace = TRUE, prob=c(0.3,0.6,0.1))),
+           Response = rbinom(200, 1, prob = 0.2))
+df
+model1<-glm(Response ~ Gender+Age_Group, data = df, family = binomial("logit"))
+summary(model1)
+
+df_agg<-df%>%group_by(Gender, Age_Group)%>%summarise(Impressions=n(), Responses=sum(Response))%>%
+  ungroup()%>%mutate(RR=Responses/Impressions)
+df_agg
+
+df_agg$No_Responses <- df_agg$Impressions- df_agg$Responses
+df_agg
+
+m3<-glm(cbind(Responses, No_Responses) ~ Gender+Age_Group, data=df_agg, family = binomial("logit"))
+summary(m3)

PCV_analysis_with_pharma.Rmd

@@ -11,16 +11,49 @@ knitr::opts_chunk$set(echo = TRUE)
 ```{r}
 library(table1)
 library(dplyr)
+library(ggplot2)
+library(GGally)
 ```
 
 ```{r}
 ### Load the data: 
 core_table <- read.csv("./Data/core_table_filled.csv")
 core_table_norace <- read.csv("./Data/core_table_droprace_filled.csv")
-supp_table_filled <- read.csv("./Data/suppl_table_filled.csv")
+supp_table_filled <- read.csv("~/Documents/Data_pharma/suppl_table_filled.csv")
 supp_table_6mo_filled <- read.csv("./Data/suppl_6mo_table_filled.csv")
 ```
 
+## Check that individuals are the same across dataframes
+```{r}
+print(paste0("Total number of individuals: "))
+sum(core_table$count)
+print("\n")
+sum(core_table_norace$count)
+print("\n")
+sum(supp_table_filled$count)
+print("\n")
+sum(supp_table_6mo_filled$count)
+```
+
+## Exploratory analysis: descriptive stats
+```{r}
+Tot_ind <- sum(supp_table_filled$count)
+vax_by_age<-supp_table_filled %>% group_by(age,PCV_combined) %>% summarise(prop = sum(count)/Tot_ind)
+vax_by_com<-supp_table_filled %>% group_by(comorbidities,PCV_combined) %>% summarise(prop = sum(count)/Tot_ind)
+
+```
+
+## Plot histograms
+```{r}
+ggplot(vax_by_age,aes(age,prop,fill=PCV_combined))+
+  geom_bar(stat="identity",position='dodge')
+ggplot(vax_by_com,aes(comorbidities,prop,fill=PCV_combined))+
+  geom_bar(stat="identity",position='dodge')
+```
+
+
+
+
 ## First step is to create another columnn of no response
 ```{r}
 ### Prepare the data and do some exploration
@@ -35,7 +68,6 @@ data_comb<-merge(data1,data2,by=c("year_month","age","patient_gender_code","race
 
 ```{r}
 ## Exploratory analysis of the dataset: 
-print(paste0("Total number of individuals: ",sum(data$count)))
 print(paste0("PCV13 vaccine coverage: ",sum(data$count[data$PCV_combined=="PCV13_only"])/sum(data$count)))
 print(paste0("PPSV23 vaccine coverage: ",sum(data$count[data$PCV_combined=="PPSV23_10yrs"])/sum(data$count)))
 print(paste0("PCV13 and PPSV23 vaccine coverage: ",sum(data$count[data$PCV_combined=="PCV13_PPSV23_10yrs"])/sum(data$count)))
@@ -58,6 +90,14 @@ m1<-glm(cbind(count.x, count.y) ~ age+patient_gender_code+race_code+PCV_combined
 #summary(m1)
 m1.ci<-confint(m1)
 ```
+##Check collinearity among the covariates using pair-wise correlation matrix
+```{r}
+X<-data_comb[,2:10]
+ggpairs(X)
+
+```
+
+
 
 ```{r}
 m1.table <- cbind(coef(m1),m1.ci)

Tests.R

@@ -0,0 +1,8 @@
+data[data$year_month=="2021-12" & data$age=="1" & data$patient_gender_code=="F"&data$race_code=="U"&data$PCV_combined=="U"&data$flu_vacc
+     =="FALSE"&data$zoster_vacc=="FALSE"&data$bmi_30_plus=="FALSE"&data$comorbidities=="FALSE"&data$income_est_mod=="U",]
+
+
+supp_table_filled[supp_table_filled$year_month=="2021-12" & supp_table_filled$age=="1" & supp_table_filled$patient_gender_code=="F"& supp_table_filled$race_code=="U"&supp_table_filled$PCV_combined=="U"&
+                    supp_table_filled$flu_vacc
+     =="FALSE"&supp_table_filled$zoster_vacc=="FALSE"&supp_table_filled$bmi_30_plus=="FALSE"&supp_table_filled$comorbidities=="FALSE"&supp_table_filled$income_est_mod=="U"
+     &supp_table_filled$COVID_severity!="non_resp_severe"& supp_table_filled$COVID_severity!="ICU_crit_care",]