Permalink
Branch: master
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
144 lines (105 sloc) 8.77 KB

Project Status: Active – The project has reached a stable, usable state and is being actively developed. Build Status DOI codecov

Loading SUPPlementary DATA into R

William D. Pearse and Scott Chamberlain

Overview

The aim of this package is to aid downloading data from published papers. To download the supplementary data from a PLoS paper, for example, you would simply type:

library(suppdata)
suppdata("10.1371/journal.pone.0127900", 1)

...and this would download the first supplementary information (SI) from the paper.

This sort of thing is very useful if you're doing meta-analyses, or just want to make sure that you know where all your data came from and want a completely reproducible "audit trail" of what you've done. It uses rcrossref to lookup which journal the article is in.

How to install and load this development version

library(devtools)
install_github("ropensci/suppdata")
library(suppdata)

This package depends on the packages httr, xml2, jsonlite, and rcrossref.

Supported publishers and repositories

See a list of potential sources at #2 - requests welcome!

Contributing

For more details on how to contribute to the package, check out the guide in CONTRIBUTING.MD.

A more detailed set of motivations for suppdata

suppdata is an R package to provide easy, reproducible access to supplemental materials within R. Thus suppdata facilitates open, reproducible research workflows: scientists re-analyzing published datasets can work with them as easily as if they were stored on their own computer, and others can track their analysis workflow painlessly.

For example, imagine you were conducting an analysis of the evolution of body mass in mammals. Without suppdata, such an analysis would require manually downloading body mass and phylogenetic data from published manuscripts. This is time-consuming, difficult (if not impossible) to make truly reproducible without re-distributing the data, and hard to follow. With suppdata, such an analysis is straightforward, reproducible, and the sources of the data are clear because their DOIs are embedded within the code:

# Load phylogenetics packages
library(ape)
library(caper)
library(phytools)

# Load suppdata
library(suppdata)

# Load two published datasets
tree <- read.nexus(suppdata("10.1111/j.1461-0248.2009.01307.x", 1))[[1]]
traits <- read.delim(suppdata("E090-184", "PanTHERIA_1-0_WR05_Aug2008.txt", "esa_archives"))

# Merge datasets
traits <- with(traits, data.frame(body.mass = log10(X5.1_AdultBodyMass_g), species=gsub(" ","_",MSW05_Binomial)))
c.data <- comparative.data(tree, traits, species)

# Calculate phylogenetic signal
phylosig(c.data$phy, c.data$data$body.mass)

A guided walk through suppdata

The aim of suppdata is to make it as easy as possible for you to write reproducible analysis scripts that make use of published data. So let's start with that first, simplest case: how to make use of published data in an analysis.

Learning by example

Below is an example of an analysis run using suppdata. Read through it first, and then we'll go through what all the parts mean.

# Load phylogenetics packages
library(ape)
library(caper)
library(phytools)

###############################
# LOAD TWO PUBLISHED DATASETS #
#       USING SUPPDATA        #
###############################
library(suppdata)
tree <- read.nexus(suppdata("10.1111/j.1461-0248.2009.01307.x", 1))[[1]]
traits <- read.delim(suppdata("E090-184", "PanTHERIA_1-0_WR05_Aug2008.txt", "esa_archives"))

# Merge datasets
traits <- with(traits, data.frame(body.mass = log10(X5.1_AdultBodyMass_g), species=gsub(" ","_",MSW05_Binomial)))
c.data <- comparative.data(tree, traits, species)

# Calculate phylogenetic signal
phylosig(c.data$phy, c.data$data$body.mass)

This short script loads some R packages focused on modelling the evolution of species' traits, then it gets to the "good stuff": using suppdata. First, we load the suppdata package using library(suppdata). The next line uses a function called read.nexus, which loads something called a phylogeny (you might be familiar with this if you're a biologist). Normally, this function would take the location of a file on our hard-drive as a single argument, but now we're giving it the output from a call to the suppdata function.

suppdata is going to the website of the article whose DOI is 10.1111/j.1461-0248.2009.01307.x (it's this paper by Fritz et al.), and then taking the first (1) supplement from that article. It saves that to a temporary location on your hard-drive, and then gives that location to read.nexus. This works with any function that expects a file at a location on your hard-drive. What particularly neat is that suppdata remembers that it has downloaded that file already (see below for more details), such that you only have to download something once per R session.

The second call to suppdata, which makes use of read.delim, shows two of the potential complexities of suppdata. First of all, because some journal publishers store their supplementary materials using numbers and others using specific file-names, suppdata takes either a number (like in the first example), or a name (like in the second example) depending on the journal publisher you're taking data from. If you look in the help file for suppdata, there is a table outlining those options. Sorry, you've just got to read up on it :-( Secondly, if you're an ecologist you might be familiar with the Ecological Society of America's data archives section. While they've moved over to a new way of storing data more recently, if you're hoping to load an older dataset from that journal you need to give the ESA data archive reference and specify that you're downloading from ESA (as in this example). If you're not an ecologist, don't worry about it, as this doesn't apply to you :D

That's it! You now know all you need to in order to use suppdata! The rest of the lines of code merge these datasets together, and then calculate something called phylogenetic signal in these datasets. If you're an evolutionary biologist, those lines might be interesting to you. If you're not, then don't worry about them.

Caching and saving to a specific directory

Sometimes, you will want to use suppdata to build a store of files on your hard-drive. If so, you should know that suppdata takes three optional arguments: cache, dir, and save.name. If you specify cache=FALSE, it will turn off suppdata's caching of files: this will force it to download your data again. This is mostly useful if you somehow make suppdata foul itself up (maybe you hit control-c or stop during a download) and so suppdata has only half-downloaded a file, and so thinks it's cached something when it hasn't. If you get an error when using suppdata, this is a good thing to try setting.

dir specifies a directory where suppdata should store files, and save.name specifies the name that the file should be saved under when saved. This is useful if you want to make a folder on your computer that contains certain files you use a lot: suppdata will cache from this folder if you tell it to, and so you can build up a reproducible selection of data to use inbetween R sessions.

rofooter