nsrr

Travis build status AppVeyor Build Status Codecov test coverage

The goal of nsrr is to allow users to access data from the National Sleep Research Resource (’NSRR’) (https://sleepdata.org/) through an R interface.

Why R? Many packages in R can read and process accelerometry, such as the GGIR, ActivityIndex, and accelerometry; other packages such as edfReader can read in EDF data for polysomnography data. Also, the xml2 package can easily read in XML annotations into R. We believe the interplay with these packages, amongst others, allow for a flexible framework to download, process, and visualize data. The nsrr package is simply the entry point into navigating the files available and downloading the data.

Installation

You can install the released version of nsrr from CRAN with:

install.packages("nsrr")

Token

To retrieve your NSRR token, go to https://sleepdata.org/dashboard, and sign in. This token will allow you access to any data sets you have requested access to. If you do not have access, then it will allow you to download files that are publicly available.

Set the token by adding this to your ~/.Renviron file:

NSRR_TOKEN="YOUR TOKEN GOES HERE"

The token is accessible via token = Sys.getenv("NSRR_TOKEN"). Each nsrr function also has the argument token to pass through if you do not wish to set it.

To determine if you are authenticated, you can use:

library(nsrr)
nsrr_auth()
$authenticated
[1] TRUE

$username
[1] "muschellij2"

$full_name
[1] ""

$first_name
NULL

$last_name
NULL

$email
[1] "muschellij2@gmail.com"

Examples

NSRR data sets

Here is how you can access the NSRR datasets list:

library(nsrr)
df = nsrr_datasets()
DT::datatable(df)

NSRR data set files

Here we first get a list of the files in the datasets sub-directory from the shhs data set:

df = nsrr_dataset_files("shhs", path = "datasets")
head(df)
  dataset                         full_path    folder
1    shhs                  datasets/archive datasets/
2    shhs             datasets/hrv-analysis datasets/
3    shhs             datasets/CHANGELOG.md datasets/
4    shhs           datasets/KNOWNISSUES.md datasets/
5    shhs datasets/shhs1-dataset-0.15.0.csv datasets/
6    shhs datasets/shhs2-dataset-0.15.0.csv datasets/
                 file_name is_file file_size
1                  archive   FALSE         0
2             hrv-analysis   FALSE         0
3             CHANGELOG.md    TRUE     11010
4           KNOWNISSUES.md    TRUE     11421
5 shhs1-dataset-0.15.0.csv    TRUE  24322962
6 shhs2-dataset-0.15.0.csv    TRUE  11897377
                 file_checksum_md5 archived
1                             <NA>    FALSE
2                             <NA>    FALSE
3 69bb54a32cdfc7bdddc13276b7c858c1    FALSE
4 8598129123baa60e16977dc24aa780af    FALSE
5 3f26d37ec97e2bc88776850a31715398    FALSE
6 21807d854010f036fd0d4f006eeed49d    FALSE

Downloading NSRR data set files

We can then download the CHANGELOG.md file as it’s publicly accessible.

url = nsrr_download_url("shhs", path = "datasets/CHANGELOG.md")
# print URL
dl = nsrr_download_file("shhs", path = "datasets/CHANGELOG.md")
dl$outfile
[1] "/var/folders/1s/wrtqcpxn685_zk570bnx9_rr0000gr/T//RtmpJUlkmk/filea17e12cd407.md"
cat(head(readLines(dl$outfile)), sep = "\n")
## 0.15.0 (November 18, 2019)

- Remove EEG spectral summary variables
- Add notes to height and weight varibles about top and bottom coding
- The CSV datasets generated from a SAS export are located here:
  - `\\rfawin\bwh-sleepepi-shhs\nsrr-prep\_releases\0.15.0\`

Listing All NSRR data set files

To list all the files, recursively, you would run:

nsrr_all_dataset_files("shhs")

but it may take some time.

EDF files

The edfReader can read in EDF fields for polysomnography data. Work in Progress - need access to EDF data.

files = nsrr_dataset_files("shhs", path= "polysomnography/edfs/shhs1")
dl = nsrr_download_file("shhs", path = "polysomnography/edfs/shhs1/shhs1-200001.edf")