blandr carries out Bland Altman analyses (also known as aTukey mean-difference plot) as described by JM Bland and DG Altman in1986. This package was started in 2015 as existing Bland-Altman Rfunctions did not calculate confidence intervals.blandr hasbeen improved iteratively over time, and hopefully will be useful as anopen-source library for reproducible science.
You can find the citation information through the usual R citationcommands:
citation("blandr")The DOI will refer to all versions of blandr. If you need to citespecific releases DOIs, the full versioning information can be found atZenodo (https://zenodo.org/record/824514), with the full sourcecode at the blandr GitHub page (https://github.com/deepankardatta/blandr/).
blandr is available as a package from CRAN and can be installed withthe following commands:
install.packages("blandr")library(blandr)You can install the blandr development version, hosted on github athttps://github.com/deepankardatta/blandr/, with thefollowing commands:
install.packages("devtools")devtools::install_github("deepankardatta/blandr")This is a basic example which shows you how to solve a commonproblem:
library(blandr)load(file="Data/bland.altman.PEFR.1986.rda")blandr.display ( bland.altman.PEFR.1986$WrightFirst , bland.altman.PEFR.1986$MiniWrightFirst ,sig.level=0.95 )blandr.draw( bland.altman.PEFR.1986$WrightFirst , bland.altman.PEFR.1986$MiniWrightFirst )One of the benefits ofblandr is that the library has beendeveloped so a module can be generated for the Jamovi GUI statisticalspreadsheet (https://www.jamovi.org/). I’m a believer in making thistool more accesible and Jamovi is a way to do this. The Jamovi modulewill also be iteratively developed - the code for each Jamovi version islinked to the CRAN version on GitHub. The module will be on the Jamovilibrary.
If you want to use the development versions ofblandr, youcan install Jamovi from the comman line using the followingcommands:
install.packages('jmvtools',repos=c('https://repo.jamovi.org','https://cran.r-project.org'))library(jmvtools)jmvtools::install()You can then download theblandr GitHub development branchand install it to Jamovi from the command line.
Thinking of a unique name was difficult - if you look through thehistory of the sources the package had a very cumbersome name. This hadto change when uploading to CRAN as it seemed to be frowned upon.
Hadley Wickham has an excellent set of documentation on creatingpackages. The one on naming (http://r-pkgs.had.co.nz/package.html) is worth a read.Reflecting on it a lot of the naming issues are to prevent collisionswith similarly named packages, so usingblandaltman in the namemight have cause problems. Mirroring the pre-existingplyr andknitr I thought I’d just add a “r” to “bland”: yes it doesn’tinvolve Altman’s name, but I couldn’t think of anything better.
If you can think of a better name please let me know!
Further information can be found in the function help files in thepackage, as well as in the vignettes.
The library is quite small, and mostly depends on the ggplot2library. However the report functions do require rmarkdown, knitr, andmost importantly pandoc - the function should check this is availableand stop gracefully if not. Further the Jamovi based functions need theJamovi library to work.
Whilst this package is good enough for use, there is the scope foriterative improvements.
Future works include:
Comments on the code, suggestions for improvement, verificationtests, validation studies, and even code contributions would begratefully accepted. Email me at deepankar(dot)datta(at)gmail.com andI’ll try and get back to you as soon as possible. Please do bear in mindthis is a side-project, and I can be otherwise busy with a lot of otherwork.