Tutorial

Use the neonUtilities Package to Access NEON Data

Authors: Claire K. Lunch, Megan A. Jones

Last Updated: Sep 23, 2024

This tutorial provides an overview of functions in the neonUtilities package in R and the neonutilities package in Python. These packages provide a toolbox of basic functionality for working with NEON data.

This tutorial is primarily an index of functions and their inputs; for more in-depth guidance in using these functions to work with NEON data, see the Download and Explore tutorial. If you are already familiar with the neonUtilities package, and need a quick reference guide to function inputs and notation, see the neonUtilities cheat sheet.

Function index

The neonUtilities/neonutilities package contains several functions (use the R and Python tabs to see the syntax in each language):

R

  • stackByTable(): Takes zip files downloaded from the Data Portal or downloaded by zipsByProduct(), unzips them, and joins the monthly files by data table to create a single file per table.
  • zipsByProduct(): A wrapper for the NEON API; downloads data based on data product and site criteria. Stores downloaded data in a format that can then be joined by stackByTable().
  • loadByProduct(): Combines the functionality of zipsByProduct(),
    stackByTable(), and readTableNEON(): Downloads the specified data, stacks the files, and loads the files to the R environment.
  • byFileAOP(): A wrapper for the NEON API; downloads remote sensing data based on data product, site, and year criteria. Preserves the file structure of the original data.
  • byTileAOP(): Downloads remote sensing data for the specified data product, subset to tiles that intersect a list of coordinates.
  • readTableNEON(): Reads NEON data tables into R, using the variables file to assign R classes to each column.
  • getCitation(): Get a BibTeX citation for a particular data product and release.

Python

  • stack_by_table(): Takes zip files downloaded from the Data Portal or downloaded by zips_by_product(), unzips them, and joins the monthly files by data table to create a single file per table.
  • zips_by_product(): A wrapper for the NEON API; downloads data based on data product and site criteria. Stores downloaded data in a format that can then be joined by stack_by_table().
  • load_by_product(): Combines the functionality of zips_by_product(),
    stack_by_table(), and read_table_neon(): Downloads the specified data, stacks the files, and loads the files to the R environment.
  • by_file_aop(): A wrapper for the NEON API; downloads remote sensing data based on data product, site, and year criteria. Preserves the file structure of the original data.
  • by_tile_aop(): Downloads remote sensing data for the specified data product, subset to tiles that intersect a list of coordinates.
  • read_table_neon(): Reads NEON data tables into R, using the variables file to assign R classes to each column.
  • get_citation(): Get a BibTeX citation for a particular data product and release.

If you are only interested in joining data files downloaded from the NEON Data Portal, you will only need to use stackByTable(). Follow the instructions in the first section of the Download and Explore tutorial.

Install and load packages

First, install and load the package. The installation step only needs to be run once, and then periodically to update when new package versions are released. The load step needs to be run every time you run your code.

R

## 
## # install neonUtilities - can skip if already installed
## install.packages("neonUtilities")
## 
## # load neonUtilities
library(neonUtilities)
##

Python

# install neonutilities - can skip if already installed
# do this in the command line
pip install neonutilities

# load neonutilities in working environment
import neonutilities as nu

Download files and load to working environment

The most popular function in neonUtilities is loadByProduct() (or load_by_product() in neonutilities). This function downloads data from the NEON API, merges the site-by-month files, and loads the resulting data tables into the programming environment, classifying each variable’s data type appropriately. It combines the actions of the zipsByProduct(), stackByTable(), and readTableNEON() functions, described below.

This is a popular choice because it ensures you’re always working with the latest data, and it ends with ready-to-use tables. However, if you use it in a workflow you run repeatedly, keep in mind it will re-download the data every time.

loadByProduct() works on most observational (OS) and sensor (IS) data, but not on surface-atmosphere exchange (SAE) data, remote sensing (AOP) data, and some of the data tables in the microbial data products. For functions that download AOP data, see the byFileAOP() and byTileAOP() sections in this tutorial. For functions that work with SAE data, see the NEON eddy flux data tutorial. SAE functions are not yet available in Python.

The inputs to loadByProduct() control which data to download and how to manage the processing:

R

  • dpID: The data product ID, e.g. DP1.00002.001
  • site: Defaults to “all”, meaning all sites with available data; can be a vector of 4-letter NEON site codes, e.g.  c("HARV","CPER","ABBY").
  • startdate and enddate: Defaults to NA, meaning all dates with available data; or a date in the form YYYY-MM, e.g.  2017-06. Since NEON data are provided in month packages, finer scale querying is not available. Both start and end date are inclusive.
  • package: Either basic or expanded data package. Expanded data packages generally include additional information about data quality, such as chemical standards and quality flags. Not every data product has an expanded package; if the expanded package is requested but there isn’t one, the basic package will be downloaded.
  • timeIndex: Defaults to “all”, to download all data; or the number of minutes in the averaging interval. See example below; only applicable to IS data.
  • release: Specify a particular data Release, e.g.  "RELEASE-2024". Defaults to the most recent Release. For more details and guidance, see the Release and Provisional tutorial.
  • include.provisional: T or F: Should provisional data be downloaded? If release is not specified, set to T to include provisional data in the download. Defaults to F.
  • savepath: the file path you want to download to; defaults to the working directory.
  • check.size: T or F: should the function pause before downloading data and warn you about the size of your download? Defaults to T; if you are using this function within a script or batch process you will want to set it to F.
  • token: Optional API token for faster downloads. See the API token tutorial.
  • nCores: Number of cores to use for parallel processing. Defaults to 1, i.e. no parallelization.

Python

  • dpid: the data product ID, e.g. DP1.00002.001
  • site: defaults to “all”, meaning all sites with available data; can be a list of 4-letter NEON site codes, e.g.  ["HARV","CPER","ABBY"].
  • startdate and enddate: defaults to NA, meaning all dates with available data; or a date in the form YYYY-MM, e.g.  2017-06. Since NEON data are provided in month packages, finer scale querying is not available. Both start and end date are inclusive.
  • package: either basic or expanded data package. Expanded data packages generally include additional information about data quality, such as chemical standards and quality flags. Not every data product has an expanded package; if the expanded package is requested but there isn’t one, the basic package will be downloaded.
  • timeindex: defaults to “all”, to download all data; or the number of minutes in the averaging interval. See example below; only applicable to IS data.
  • release: Specify a particular data Release, e.g.  "RELEASE-2024". Defaults to the most recent Release. For more details and guidance, see the Release and Provisional tutorial.
  • include_provisional: True or False: Should provisional data be downloaded? If release is not specified, set to T to include provisional data in the download. Defaults to F.
  • savepath: the file path you want to download to; defaults to the working directory.
  • check_size: True or False: should the function pause before downloading data and warn you about the size of your download? Defaults to True; if you are using this function within a script or batch process you will want to set it to False.
  • token: Optional API token for faster downloads. See the API token tutorial.
  • cloud_mode: Can be set to True if you are working in a cloud environment; provides more efficient data transfer from NEON cloud storage to other cloud environments.
  • progress: Set to False to omit the progress bar during download and stacking.

The dpID (dpid) is the data product identifier of the data you want to download. The DPID can be found on the Explore Data Products page. It will be in the form DP#.#####.###

Demo data download and read

Let’s get triple-aspirated air temperature data (DP1.00003.001) from Moab and Onaqui (MOAB and ONAQ), from May–August 2018, and name the data object triptemp:

R

triptemp <- loadByProduct(dpID="DP1.00003.001", 
                          site=c("MOAB","ONAQ"),
                          startdate="2018-05", 
                          enddate="2018-08")

Python

triptemp = nu.load_by_product(dpid="DP1.00003.001", 
                              site=["MOAB","ONAQ"],
                              startdate="2018-05", 
                              enddate="2018-08")

View downloaded data

The object returned by loadByProduct() is a named list of data tables, or a dictionary of data tables in Python. To work with each of them, select them from the list.

R

names(triptemp)
## [1] "citation_00003_RELEASE-2024" "issueLog_00003"             
## [3] "readme_00003"                "sensor_positions_00003"     
## [5] "TAAT_1min"                   "TAAT_30min"                 
## [7] "variables_00003"
temp30 <- triptemp$TAAT_30min

If you prefer to extract each table from the list and work with it as an independent object, you can use the list2env() function:

list2env(trip.temp, .GlobalEnv)

Python

triptemp.keys()
## dict_keys(['TAAT_1min', 'TAAT_30min', 'citation_00003_RELEASE-2024', 'issueLog_00003', 'readme_00003', 'sensor_positions_00003', 'variables_00003'])
temp30 = triptemp["TAAT_30min"]

If you prefer to extract each table from the list and work with it as an independent object, you can use
globals().update():

globals().update(triptemp)

For more details about the contents of the data tables and metadata tables, check out the Download and Explore tutorial.

Join data files: stackByTable()

The function stackByTable() joins the month-by-site files from a data download. The output will yield data grouped into new files by table name. For example, the single aspirated air temperature data product contains 1 minute and 30 minute interval data. The output from this function is one .csv with 1 minute data and one .csv with 30 minute data.

Depending on your file size this function may run for a while. For example, in testing for this tutorial, 124 MB of temperature data took about 4 minutes to stack. A progress bar will display while the stacking is in progress.

Download the Data

To stack data from the Portal, first download the data of interest from the NEON Data Portal. To stack data downloaded from the API, see the zipsByProduct() section below.

Your data will download from the Portal in a single zipped file.

The stacking function will only work on zipped Comma Separated Value (.csv) files and not the NEON data stored in other formats (HDF5, etc).

Run stackByTable()

The example data below are single-aspirated air temperature.

To run the stackByTable() function, input the file path to the downloaded and zipped file.

R

# Modify the file path to the file location on your computer
stackByTable(filepath="~neon/data/NEON_temp-air-single.zip")

Python

# Modify the file path to the file location on your computer
nu.stack_by_table(filepath="/neon/data/NEON_temp-air-single.zip")

In the same directory as the zipped file, you should now have an unzipped directory of the same name. When you open this you will see a new directory called stackedFiles. This directory contains one or more .csv files (depends on the data product you are working with) with all the data from the months & sites you downloaded. There will also be a single copy of the associated variables, validation, and sensor_positions files, if applicable (validation files are only available for observational data products, and sensor position files are only available for instrument data products).

These .csv files are now ready for use with the program of your choice.

To read the data tables, we recommend using readTableNEON(), which will assign each column to the appropriate data type, based on the metadata in the variables file. This ensures time stamps and missing data are interpreted correctly.

Load data to environment

R

SAAT30 <- readTableNEON(
  dataFile='~/stackedFiles/SAAT_30min.csv',
  varFile='~/stackedFiles/variables_00002.csv'
)

Python

SAAT30 = nu.read_table_neon(
  dataFile='/stackedFiles/SAAT_30min.csv',
  varFile='/stackedFiles/variables_00002.csv'
)

Other function inputs

Other input options in stackByTable() are:

  • savepath : allows you to specify the file path where you want the stacked files to go, overriding the default. Set to "envt" to load the files to the working environment.
  • saveUnzippedFiles : allows you to keep the unzipped, unstacked files from an intermediate stage of the process; by default they are discarded.

Example usage:

R

stackByTable(filepath="~neon/data/NEON_temp-air-single.zip", 
             savepath="~data/allTemperature", saveUnzippedFiles=T)

tempsing <- stackByTable(filepath="~neon/data/NEON_temp-air-single.zip",
savepath="envt", saveUnzippedFiles=F)

Python

nu.stack_by_table(filepath="/neon/data/NEON_temp-air-single.zip", 
                  savepath="/data/allTemperature", 
                  saveUnzippedFiles=True)

tempsing <- nu.stack_by_table(filepath="/neon/data/NEON_temp-air-single.zip",
savepath="envt",
saveUnzippedFiles=False)

Download files to be stacked: zipsByProduct()

The function zipsByProduct() is a wrapper for the NEON API, it downloads zip files for the data product specified and stores them in a format that can then be passed on to stackByTable().

Input options for zipsByProduct() are the same as those for loadByProduct() described above.

Here, we’ll download single-aspirated air temperature (DP1.00002.001) data from Wind River Experimental Forest (WREF) for April and May of 2019.

R

zipsByProduct(dpID="DP1.00002.001", site="WREF", 
              startdate="2019-04", enddate="2019-05",
              package="basic", check.size=T)

Downloaded files can now be passed to stackByTable() to be stacked.

stackByTable(filepath=paste(getwd(), 
                            "/filesToStack00002", 
                            sep=""))

Python

nu.zips_by_product(dpid="DP1.00002.001", site="WREF", 
                   startdate="2019-04", enddate="2019-05",
                   package="basic", check_size=True)

Downloaded files can now be passed to stackByTable() to be stacked.

nu.stack_by_table(filepath=os.getcwd()+
                  "/filesToStack00002")

For many sensor data products, download sizes can get very large, and stackByTable() takes a long time. The 1-minute or 2-minute files are much larger than the longer averaging intervals, so if you don’t need high- frequency data, the timeIndex input option lets you choose which averaging interval to download.

This option is only applicable to sensor (IS) data, since OS data are not averaged.

Download by averaging interval

Download only the 30-minute data for single-aspirated air temperature at WREF:

R

zipsByProduct(dpID="DP1.00002.001", site="WREF", 
              startdate="2019-04", enddate="2019-05",
              package="basic", timeIndex=30, 
              check.size=T)

Python

nu.zips_by_product(dpid="DP1.00002.001", site="WREF", 
                   startdate="2019-04", 
                   enddate="2019-05", package="basic", 
                   timeindex=30, check_size=True)

The 30-minute files can be stacked and loaded as usual.

Download remote sensing files

Remote sensing data files can be very large, and NEON remote sensing (AOP) data are stored in a directory structure that makes them easier to navigate. byFileAOP() downloads AOP files from the API while preserving their directory structure. This provides a convenient way to access AOP data programmatically.

Be aware that downloads from byFileAOP() can take a VERY long time, depending on the data you request and your connection speed. You may need to run the function and then leave your machine on and downloading for an extended period of time.

Here the example download is the Ecosystem Structure data product at Hop Brook (HOPB) in 2017; we use this as the example because it’s a relatively small year-site-product combination.

R

byFileAOP("DP3.30015.001", site="HOPB", 
          year=2017, check.size=T)

Python

nu.by_file_aop(dpid="DP3.30015.001", 
               site="HOPB", year=2017, 
               check_size=True)

The files should now be downloaded to a new folder in your working directory.

Download remote sensing files for specific coordinates

Often when using remote sensing data, we only want data covering a certain area - usually the area where we have coordinated ground sampling. byTileAOP() queries for data tiles containing a specified list of coordinates. It only works for the tiled, AKA mosaicked, versions of the remote sensing data, i.e. the ones with data product IDs beginning with “DP3”.

Here, we’ll download tiles of vegetation indices data (DP3.30026.001) corresponding to select observational sampling plots. For more information about accessing NEON spatial data, see the API tutorial and the in-development geoNEON package.

For now, assume we’ve used the API to look up the plot centroids of plots SOAP_009 and SOAP_011 at the Soaproot Saddle site. You can also look these up in the Spatial Data folder of the document library. The coordinates of the two plots in UTMs are 298755,4101405 and 299296,4101461. These are 40x40m plots, so in looking for tiles that contain the plots, we want to include a 20m buffer. The “buffer” is actually a square, it’s a delta applied equally to both the easting and northing coordinates.

R

byTileAOP(dpID="DP3.30026.001", site="SOAP", 
          year=2018, easting=c(298755,299296),
          northing=c(4101405,4101461),
          buffer=20)

Python

nu.by_tile_aop(dpid="DP3.30026.001", 
               site="SOAP", year=2018, 
               easting=[298755,299296],
               northing=[4101405,4101461],
               buffer=20)

The 2 tiles covering the SOAP_009 and SOAP_011 plots have
been downloaded.

Questions?

If you have questions or comments on this content, please contact us.

Contact Us