intubate (version 1.0.0)

survey: Interfaces for survey package for data science pipelines.

Description

Interfaces to survey functions that can be used in a pipeline implemented by magrittr.

Usage

ntbt_svyby(data, ...)         ## data <-> design
ntbt_svycoxph(data, ...)      ## data <-> design
ntbt_svydesign(data, ...)
ntbt_svyglm(data, ...)        ## data <-> design
ntbt_svymean(data, ...)       ## data <-> design
ntbt_svyquantile(data, ...)   ## data <-> design
ntbt_svyratio(data, ...)      ## data <-> design
ntbt_svytotal(data, ...)      ## data <-> design
ntbt_twophase(data, ...)

Arguments

data

data frame, tibble, list, ...

...

Other arguments passed to the corresponding interfaced function.

Value

Object returned by interfaced function.

Details

Interfaces call their corresponding interfaced function.

Examples

## Not run: 
# library(intubate)
# library(magrittr)
# library(survey)
# 
# ## svydesign
# data(api)
# ## Original function to interface
# # stratified sample
# dstrat <- svydesign(id=~1,strata=~stype, weights=~pw, data=apistrat, fpc=~fpc)
# # one-stage cluster sample
# dclus1 <- svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)
# # two-stage cluster sample: weights computed from population sizes.
# dclus2 <- svydesign(id=~dnum+snum, fpc=~fpc1+fpc2, data=apiclus2)
# 
# ## The interface puts data as first parameter
# # stratified sample
# dstrat <- ntbt_svydesign(data=apistrat, id=~1,strata=~stype, weights=~pw, fpc=~fpc)
# # one-stage cluster sample
# dclus1 <- ntbt_svydesign(data=apiclus1, id=~dnum, weights=~pw, fpc=~fpc)
# # two-stage cluster sample: weights computed from population sizes.
# dclus2 <- ntbt_svydesign(data=apiclus2, id=~dnum+snum, fpc=~fpc1+fpc2)
# 
# ## so it can be used easily in a pipeline.
# dstrat <- apistrat %>%
#   ntbt_svydesign(id=~1,strata=~stype, weights=~pw, fpc=~fpc)
# # one-stage cluster sample
# dclus1 <- apiclus1 %>%
#   ntbt_svydesign(id=~dnum, weights=~pw, fpc=~fpc)
# # two-stage cluster sample: weights computed from population sizes.
# dclus2 <- apiclus2 %>%
#   ntbt_svydesign(id=~dnum+snum, fpc=~fpc1+fpc2)
# 
# ## twofase
# ## two-phase simple random sampling.
# data(pbc, package="survival")
# pbc$randomized <- with(pbc, !is.na(trt) & trt>0)
# pbc$id<-1:nrow(pbc)
# 
# ## Original function to interface
# d2pbc <- twophase(id=list(~id,~id), data=pbc, subset=~randomized)
# svymean(~bili, d2pbc)
# 
# ## The interface puts data as first parameter
# d2pbc <- ntbt_twophase(data=pbc, id=list(~id,~id), subset=~randomized)
# svymean(~bili, d2pbc)
# 
# ## so it can be used easily in a pipeline.
# d2pbc <- pbc %>%
#   ntbt_twophase(id=list(~id,~id), subset=~randomized)
# svymean(~bili, d2pbc)
# 
# 
# ## ntbt_svyby, ntbt_svyglm, ntbt_svymean,
# ## ntbt_svyquantile, ntbt_svyratio, ntbt_svytotal
# 
# ## From vignette of survey
# vars<-names(apiclus1)[c(12:13,16:23,27:37)] 
# 
# ## original
# dclus1 <- svydesign(id = ~dnum, weights = ~pw, data = apiclus1, fpc = ~fpc)
# summary(dclus1)
# ## direct call
# dclus1 <- apiclus1 %>%
#   ntbt(svydesign, id = ~dnum, weights = ~pw, fpc = ~fpc)
# summary(dclus1)
# ## interface
# dclus1 <- apiclus1 %>%
#   ntbt_svydesign(id = ~dnum, weights = ~pw, fpc = ~fpc)
# summary(dclus1)
# 
# ## original
# svymean(~api00, dclus1)
# ## direct call
# dclus1 %>%
#   ntbt(svymean, x=~api00)
# ## interface
# dclus1 %>%
#   ntbt_svymean(x=~api00)
# 
# ## original
# svyquantile(~api00, dclus1, quantile=c(0.25,0.5,0.75), ci=TRUE)
# ## direct call
# dclus1 %>%
#   ntbt(svyquantile, ~api00, quantile=c(0.25,0.5,0.75), ci=TRUE)
# ## interface
# dclus1 %>%
#   ntbt(svyquantile, ~api00, quantile=c(0.25,0.5,0.75), ci=TRUE)
# 
# ## original
# svytotal(~stype, dclus1)
# svytotal(~enroll, dclus1)
# ## direct call
# dclus1 %>%
#   ntbt(svytotal, ~stype)
# dclus1 %>%
#   ntbt(svytotal,~enroll)
# ## interface
# dclus1 %>%
#   ntbt(svytotal, ~stype)
# dclus1 %>%
#   ntbt(svytotal,~enroll)
# 
# ## original
# svyratio(~api.stu, ~enroll, dclus1)
# svyratio(~api.stu, ~enroll, design=subset(dclus1, stype=="H"))
# ## direct call
# dclus1 %>%
#   ntbt(svyratio, ~api.stu, ~enroll)
# dclus1 %>%
#   ntbt(svyratio, ~api.stu, ~enroll, design=subset("#", stype=="H"))
# ## interface
# dclus1 %>%
#   ntbt_svyratio(~api.stu, ~enroll)
# dclus1 %>%
#   ntbt_svyratio(~api.stu, ~enroll, design=subset("#", stype=="H"))
# 
# ## original
# svymean(make.formula(vars),dclus1,na.rm=TRUE)
# ## direct call
# dclus1 %>%
#   ntbt(svymean, make.formula(vars), na.rm=TRUE)
# ## interface
# dclus1 %>%
#   ntbt_svymean(make.formula(vars), na.rm=TRUE)
# 
# ## original
# svyby(~ell+meals, ~stype, design=dclus1, svymean)
# ## direct call
# dclus1 %>%
#   ntbt(svyby, ~ell+meals, ~stype, svymean)
# ## interface
# dclus1 %>%
#   ntbt_svyby(~ell+meals, ~stype, svymean)
# 
# ## original
# regmodel <- svyglm(api00~ell+meals, design=dclus1)
# summary(regmodel)
# logitmodel <- svyglm(I(sch.wide=="Yes")~ell+meals, design=dclus1,
#                      family=quasibinomial()) 
# summary(logitmodel)
# ## direct call
# dclus1 %>%
#   ntbt(svyglm, api00~ell+meals) %>%
#   summary()
# dclus1 %>%
#   ntbt(svyglm, I(sch.wide=="Yes")~ell+meals, family=quasibinomial()) %>%
#   summary()
# ## interface
# dclus1 %>%
#   ntbt_svyglm(api00~ell+meals) %>%
#   summary()
# dclus1 %>%
#   ntbt_svyglm(I(sch.wide=="Yes")~ell+meals, family=quasibinomial()) %>%
#   summary()
# 
# ## ntbt_svycoxph
# ## stratified on case status
# data(nwtco)
# ## original
# dcchs <- twophase(id=list(~seqno,~seqno), strata=list(NULL,~rel),
#                   subset=~I(in.subcohort | rel), data=nwtco)
# svycoxph(Surv(edrel,rel)~factor(stage)+factor(histol)+I(age/12), design=dcchs)
# ## direct call
# nwtco %>%
#   ntbt(twophase,id = list(~seqno,~seqno), strata = list(NULL,~rel),
#        subset = ~I(in.subcohort | rel)) %>%
#   ntbt(svycoxph, Surv(edrel,rel)~factor(stage)+factor(histol)+I(age/12))
# ## interface
# nwtco %>%
#   ntbt_twophase(id = list(~seqno,~seqno), strata = list(NULL,~rel),
#        subset = ~I(in.subcohort | rel)) %>%
#   ntbt_svycoxph(Surv(edrel,rel)~factor(stage)+factor(histol)+I(age/12))
# 
# ## Involved example using `intubOrders`, transforming the code in:
# 
# ## https://cran.r-project.org/web/packages/survey/vignettes/survey.pdf
# 
# data(api)
# 
# ## First, the original code from the vignette
# vars<-names(apiclus1)[c(12:13,16:23,27:37)] 
# 
# dclus1 <- svydesign(id = ~dnum, weights = ~pw, data = apiclus1, fpc = ~fpc)
# summary(dclus1)
# svymean(~api00, dclus1)
# svyquantile(~api00, dclus1, quantile=c(0.25,0.5,0.75), ci=TRUE)
# svytotal(~stype, dclus1)
# svytotal(~enroll, dclus1)
# svyratio(~api.stu,~enroll, dclus1)
# svyratio(~api.stu, ~enroll, design=subset(dclus1, stype=="H"))
# svymean(make.formula(vars),dclus1,na.rm=TRUE)
# svyby(~ell+meals, ~stype, design=dclus1, svymean)
# regmodel <- svyglm(api00~ell+meals,design=dclus1)
# logitmodel <- svyglm(I(sch.wide=="Yes")~ell+meals, design=dclus1, family=quasibinomial()) 
# summary(regmodel)
# summary(logitmodel)
# 
# ## Now using intubOrders and ntbt.
# 
# ## Strategy 1: long pipeline, light use of intubOrders.
# 
# apiclus1 %>%
#   ntbt(svydesign, id = ~dnum, weights = ~ pw, fpc = ~ fpc, "<|| summary >") %>%
#   ntbt(svymean, ~ api00, "<|f| print >") %>%
#   ntbt(svyquantile, ~ api00, quantile = c(0.25,0.5,0.75), ci = TRUE, "<|f| print >") %>%
#   ntbt(svytotal, ~ stype, "<|f| print >") %>%
#   ntbt(svytotal, ~ enroll, "<|f| print >") %>%
#   ntbt(svyratio, ~ api.stu, ~ enroll, "<|f| print >") %>%
#   ntbt(svyratio, ~ api.stu, ~ enroll, design = subset("#", stype == "H"), "<|f| print >") %>%
#   ntbt(svymean, make.formula(vars), na.rm = TRUE, "<|f| print >") %>%
#   ntbt(svyby, ~ ell + meals, ~ stype, svymean, "<|f| print >") %>%
#   ntbt(svyglm, api00 ~ ell + meals, "<|f| summary >") %>%
#   ntbt(svyglm, I(sch.wide == "Yes") ~ ell + meals, family = quasibinomial(), "<|f| summary >") %>%
#   summary() ## We have forwarded the result from svydesign (line 2),
#             ## so we could still continue using it downstream.
# 
# ## Strategy 2: short pipeline, heavy use of *one* intubOrder.
# apiclus1 %>%
#   ntbt(svydesign, id = ~dnum, weights = ~pw, fpc = ~fpc,
#        "<|f|
#          summary;
#          svymean(~api00, #);
#          svyquantile(~api00, #, quantile = c(0.25, 0.5, 0.75), ci = TRUE);
#          svytotal(~stype, #);
#          svytotal(~enroll, #);
#          svyratio(~api.stu,~enroll, #);
#          svyratio(~api.stu, ~enroll, design = subset(#, stype == 'H'));
#          svymean(make.formula(vars), #, na.rm = TRUE);
#          svyby(~ell+meals, ~stype, #, svymean);
#          summary(svyglm(api00~ell+meals, #));
#          summary(svyglm(I(sch.wide == 'Yes')~ell+meals, #, family = quasibinomial())) >") %>%
#   head()  ## We have forwarded the original dataset,
#           ## so we could continue using it downstream.
# ## End(Not run)