In serghiou/rtransparent: Identifies indicators of transparency

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  message = FALSE,
  warning = FALSE 
)

rtransparent can be used to analyze TXT and PMC XML files. This vignette illustrates how to access example data and use this package to identify and extract indicators of transparency.

# Load rtransparent
library(rtransparent)

Note that to run the following code you need to have installed the dplyr and metareadr packages.

TXT files

Data

First, I downloaded the PDF of the open access publication Reproducible research practices, transparency, and open access data in the biomedical literature, 2015–2017 and saved it as "PMID30457984-PMC6245499.pdf" - note that this could also be done programmatically by using the fantastic package fulltext. Now, let us convert this into a TXT file. Note that for this function to work you need to have installed the poppler PDF rendering library. This can be a pain, but the easiest way to do this is by using Homebrew.

# Extract txt from pdf
article <- rt_read_pdf("../inst/extdata/PMID32171256-PMC7071725.pdf")

# Print the first 200 characters
cat(substr(article, start = 1, stop = 200))

# Save
write(article, "PMID32171256.txt")

Note that the package automatically takes the name of the file as the PMID (PubMed ID) when reading the TXT file. As such, either name your file using its PMID (as I have done here), or disregard the column "pmid" in the resulting dataframes (see below).

Detect indicators of transparency

Search for any availability of data or code. Note that this is done by utilizing Nico Riedel's great package oddpub.

data_code <- rt_data_code("PMID32171256.txt")

# Glimpse
dplyr::glimpse(data_code)

Search for mention of Conflicts of interest (COI).

coi <- rt_coi("PMID32171256.txt")

# Glimpse
dplyr::glimpse(coi)

Search for mention of Funding.

fund <- rt_fund("PMID32171256.txt")

# Glimpse
dplyr::glimpse(fund)

Search for mention of Protocol registration.

register <- rt_register("PMID32171256.txt")

# Glimpse
dplyr::glimpse(register)

Search for all of COI, Funding and Protocol registration concurrently. Note that the functions for data and code utilize the oddpub package. As such, they use a slightly different approach to parallelization, which was necessary in running this function across millions of articles, which was not compatible with the rest of the functions. This is why the data and code functions were not implemented within the rt_all function.

all_indicators <- rt_all("PMID32171256.txt")

# Glimpse
dplyr::glimpse(all_indicators)

PMC XML files

Data

First, let us download the PMC XML file for the same article: Reproducible research practices, transparency, and open access data in the biomedical literature, 2015–2017. To do that, I will be using the metareadr package and the PMC ID (PubMed Central ID) of this article (PMC 6245499) (download this package using devtools::install_github("serghiou/metareadr").

# Download XML file
metareadr::mt_read_pmcoa(pmcid = "7071725", file_name = "PMID32171256.xml")

Detect indicators of transparency

Search for any availability of data or code. Note that this is done by utilizing Nico Riedel's package oddpub. Note also the use of the remove_ns argument to remove the PMC XML namespace - this is not required if using the new version of metareadr, but illustrated in case you are using an old version or downloading these files by yourself.

data_code <- rt_data_code_pmc("PMID32171256.xml", remove_ns = T)

# Glimpse
dplyr::glimpse(data_code)

Search for mention of Conflicts of interest (COI).

coi <- rt_coi_pmc("PMID32171256.xml", remove_ns = T)

# Glimpse
dplyr::glimpse(coi)

Search for mention of Funding.

fund <- rt_fund_pmc("PMID32171256.xml",  remove_ns = T)

# Glimpse
dplyr::glimpse(fund)

Search for mention of Protocol registration.

register <- rt_register_pmc("PMID32171256.xml",  remove_ns = T)

# Glimpse
dplyr::glimpse(register)

Search for all of COI, Funding and Protocol registration concurrently. Note that the functions for data and code utilize the oddpub package. As such, they use a slightly different approach to parallelization, which was necessary in running this function across millions of articles, which was not compatible with the rest of the functions. This is why the data and code functions were not implemented within the rt_all function.

all_indicators <- rt_all_pmc("PMID32171256.xml",  remove_ns = T)

# Glimpse
dplyr::glimpse(all_indicators)

Remove downloaded files.

file.remove("PMID30457984.txt", "PMID30457984.xml")

serghiou/rtransparent documentation built on Dec. 26, 2024, 8:19 p.m.