Data Discovery & Management

Overview

Synthesis projects often begin with a few datasets that inspire the questions–and end up incorporating dozens or hundreds of others. Researchers may seek out data that resemble their initial datasets, but come from other climates, ecosystems, or cultural settings. Or they may find that they need data of a completely different kind to establish drivers and context. The best synthesizers are resourceful in their search for data, cautious in evaluating data quality and relevance, and meticulous in documenting data sources, treatments, and related analytical decisions. In this workshop, we will cover all these aspects in enough depth for participants to begin finding and assessing their own project data.

Learning Objectives

After completing this module you will be able to:

  • Identify repositories “known for” a particular type of data
  • Explain how to effectively search for data outside of specialized repositories
  • Create a data inventory for identified data that allows for easy re-finding of those data products
  • Plan how to download data in a reproducible way
  • Perform checks of the fundamental structure of a dataset

Preparation

Each Synthesis Fellow should identify five (5) datasets of possible relevance to their project and enter the specified information into the first sheet of the “data-inventory” Google Sheet.

Panel Discussion

To motivate this module and provide some beneficial context, we’re beginning with a conversation with a panel composed of people who work at various organizations with a focus on data management and production. See the tabs below for each year’s panelists and links to their professional sites.

Panelists will briefly introduce themselves and describe their roles. They will then speak to the kinds of data available at their organization and the strengths, limitations, and quirks of those data products from a synthesis lens. Individuals not associated with data repositories will instead share their experience working with specific types of data. Time allowing, panelists will talk about their experiences working at their organizations more broadly.

  • Dr. Greg Maurer, Environmental Data Initiative (EDI) and Jornada LTER
  • Dr. Eric Sokol, Staff Scientist, Quantitative Ecology, National Ecological Observatory Network (NEON)
  • Dr. Nicole Kaplan, Computational Biologist, U.S. Department of Agriculture-Agricultural Research Service (USDA-ARS)
  • Dr. Steve Formel, Biologist, USGS Science, Analytics, and Synthesis Program and node manager for the Ocean Biodiversity Information System - USA (OBIS-USA) and the Global Biodiversity Information Facility US (GBIF-US)

Pre-Prepared Questions

  • What policies are in place to ensure responsible use of your data?
  • What challenges (technical and scientific) do you see in integrating data across platforms and organizations?
  • Are you aware of any open sources of code useful for downloading, wrangling, or analyzing data in your repository?
  • How can young scientists and data professionals contribute to the work being done by your organizations?

Data Repositories

There are a lot of specialized data repositories out there. These organizations are either primarily dedicated to storing and managing data or those operations constitute a substantive proportion of their efforts. In synthesis work, you may already have some datasets in-hand at the outset but it likely that you will need to find more data to test your hypotheses. Data repositories are a great way of finding/accessing data that are relevant to your questions.

You’ll become familiar with many of these when you need a particular type of data and go searching for it but to help speed you along, see the list below for a non-exhaustive set of some that have proved useful to other synthesis projects in the past. They are in alphabetical order. If the “ Package” column contains the GitHub logo () then the package is available on GitHub but is not available on CRAN (or not available at time of writing).

Name Description Package
AmeriFlux Provides data on carbon, water, and energy fluxes in ecosystems across the Americas, aiding in climate change and carbon cycle research. amerifluxr
DataONE Aggregates environmental and ecological data from global sources, focusing on biodiversity, climate, and ecosystem research. dataone
EDI Contains a wide range of ecological and environmental datasets, including long-term observational data, experimental results, and field studies from diverse ecosystems. EDIutils
EES-DIVE The Environmental System Science Data Infrastructure for a Virtual Ecosystem (ESS-DIVE) includes a variety of observational, experimental, modeling and other data products from a wide range of ecological and urban systems. –
GBIF The Global Biodiversity Information Facility (GBIF) aggregates global species occurrence data and biodiversity records, supporting research in species distribution and conservation. rgbif
Google Earth Engine Google Earth Engine is a cloud-based geospatial analysis platform that provides access to vast amounts of satellite imagery and environmental data for monitoring and understanding changes in the Earth’s surface. rgee
Microsoft Planetary Computer The Microsoft Planetary Computer is a cloud-based platform that combines global environmental datasets with advanced analytical tools to support sustainability and ecological research. rstac
NASA Provides data on earth science, space exploration, and climate, including satellite imagery and observational data for both terrestrial and extraterrestrial studies. Nice GUI-based data download via AppEEARS. nasadata
NCBI Hosts genomic and biological data, including DNA, RNA, and protein sequences, supporting genomics and molecular biology research. rentrez
NEON Provides ecological data from U.S. field sites, covering biodiversity, ecosystems, and environmental changes, supporting large-scale ecological research. neonUtilities
NOAA Offers meteorological, oceanographic, and climate data, essential for understanding atmospheric conditions, marine environments, and long-term climate trends. EpiNOAA-R
Open Traits Network While not a repository per se, the Open Traits Network has compiled an extensive lists of repositories for trait data. Check out their repository inventory for trait data –
USGS Hosts data on geology, hydrology, biology, and geography, including topographical maps and natural resource assessments. dataRetrieval

General Data Searches

If you don’t find what you’re looking for in a particular data repository (or want to look for data not included in one of those platforms), you might want to consider a broader search. For instance, Google is a suprisingly good resource for finding data and–for those familiar with Google Scholar for peer reviewed literature-specific Googling–there is a dataset-specific variant of Google called Google Dataset Search.

Search Operators

Virtually all search engines support “operators” to create more effective queries (i.e., search parameters). If you don’t use operators, most systems will just return results that have any of the words in your search which is non-ideal, especially when you’re looking for very specific criteria in candidate datasets.

See the tabs below for some useful operators that might help narrow your dataset search even when using more general platforms.

Use quotation marks ("") to search for an exact phrase. This is particularly useful when you need specific data points or exact wording.

Example: "reef biodiversity"

Use an asterisk (*) to search using a placeholder for any word or phrase in the query. This is useful for finding variations of a term.

Example: Pinus * data

Use a plus sign (+) to search using more than one query at the same time. This is useful when you need combinations of criteria to be met.

Example: bat + cactus

Use the ‘or’ operator (OR) operator to search for either one term or another. It broadens your search to include multiple terms.

Example: "prairie pollinator" OR "grassland pollinator"

Use a minus sign (-; a.k.a. “hyphen”) to exclude certain words from your search. Useful to filter out irrelevant results.

Example: marine biodiversity data -fishery

Use the site operator (site:) to search within a specific website or domain. This is helpful when you’re looking for data from a particular source.

Example: site:.gov bird data

Use the file type operator (filetype:) to search for data with a specific file extension. Useful to make sure the data you find is already in a format you can intteract with.

Example: filetype:tif precipitation data

Use the ‘in title’ operator (intitle:) to search for pages that have a specific word in the title. This can narrow down your results to more relevant pages.

Example: intitle:"lithology"

Use the ‘in URL’ operator (inurl:) to search for pages that have a specific word in the URL. This can help locate data repositories or specific datasets.

Example: inurl:data soil chemistry

Data Inventory Value

Documenting potential datasets (and their metadata) thoroughly in a data inventory provides numerous benefits! These include:

  • Well-documented data inventory table make it easier for researchers to find and access specific data for reproducible research
  • Documentation will help researchers to quickly understand the context, scope, and limitations of the data, reducing the time spent on preliminary data assessment
  • Detailed documentation will speed up the data publication process (e.g., data provenance, the difference among methods, etc.)
Activity: Data Inventory

Part 1 (~25 min)

In your project groups:

  • Review your data inventory Google Sheet and discuss why you chose the datasets included.
  • Discuss what key information is needed to determine if each dataset is useful for your project. The columns from the detailed data inventory sheet (Data_Inventory_V2) might help to guide the conversation.
  • Decide on the datasets you want to use in your analysis, ensuring there are at least as many as group members. Each group member will self-assign at least one dataset.
    • Later, each person will download their assigned dataset
  • After identifying the necessary information, begin filling out the detailed data inventory Google Sheet. Focus on completing the columns that are relevant for making informed decisions.
    • This sheet will be shared with another group in Part 2

Part 2 (~10 min)

  • You are assigned to review the detailed data inventory table (Data_Inventory_V2) from another project group
  • Each group member self-assigns one dataset that you want to review.
  • Try to find the exact data file to which you were assigned
  • Do you agree with the information entered in the data inventory?
  • Is there any information you think should be in the data inventory that wasn’t?
Discussion: Data Inventory

Return to the main room and let’s discuss (some of) the following questions:

  • Which elements of the data inventory table made it easier or more difficult to find the data?
  • What challenges did you encounter while searching for the datasets?
  • What is your plan for downloading the data?

Downloading Data

Once you’ve found data, filled out your data inventory, and decided which datasets you actually want, it’s time to download some of them! There are several methods you can use and it’s possible that each won’t work in all cases so it’s important to be at least somewhat familiar with several of these tools.

Most of these methods will work regardless of the format of the data (i.e., its file extension) but the format of the data will be important when you want to ‘read in’ the data and begin to work with it.

Below are some example code chunks for five methods of downloading data in a scripted way. There will be contexts where only a Graphical User Interface (“GUI”; [GOO-ee]) is available but the details of that method of downloading are usually specific to the portal you’re accessing so we won’t include an artificial general case.

Sometimes you might have a URL directly to a particular dataset (usually one hosted by a data repository). If you copy/paste this URL into your browser the download would automatically begin. However, we want to make our workflows entirely scripted (or close to it) so see the example below for how to download data via a data entity URL.

The dataset we download below is one collected at the Santa Barbara Coastal (SBC) LTER on California spiny lobster (Panulirus interruptus) populations.

# Define URL as an object
dt_url <- "https://pasta.lternet.edu/package/data/eml/knb-lter-sbc/77/10/f32823fba432f58f66c06b589b7efac6"

# Read it into R
lobster_df <- read.csv(file = dt_url)
1
You can typically find this URL in the repository where you found the dataset

If you’re quite lucky, the data you want might be stored in a repository that developed (and maintains!) an R package. These packages may or may not be on CRAN (packages can often also be found on GitHub or Bioconductor). Typically these packages have a “vignette” that demonstrates how their functions should be used.

Consider the following example adapted from the USGS dataRetrieval package vignette. Visit USGS National Water Dashboard interactive map to find site numbers and check data availability.

# Load needed packages
## install.packages("librarian")
librarian::shelf(dataRetrieval)

# Set up the parameters for the Santa Ynez River site
siteNumber <- "11133000"  # USGS site number for Santa Ynez River at Narrows
parameterCd <- "00060"    # Parameter code for discharge (cubic feet per second)
startDate <- "2024-01-01" # Start date
endDate <- "2024-01-31"   # End date

# Retrieve daily discharge data
dischargeData <- readNWISdv(siteNumber, parameterCd, startDate, endDate)

# View the first few rows of the data
head(dischargeData)

You may want to download several data files hosted in the same repository online for different spatial/temporal replicates. You could try to use the data entity URL or an associated package (if one exists) or you could write code to do a “batch download” where you’d just download each file using a piece of code that repeats itself as much as needed.

The dataset we demonstrate downloading below is NOAA weather station data. Specifically it is the Integrated Surface Data (ISD).

# Specify the start/end years for which you want to download data
target_years <- 2000:2005

# Loop across years
for(focal_year in target_years){

  # Message a progress note
  message("Downloading data for ", focal_year)

  # Assemble the URL manually
  focal_url <- paste0( "https://www1.ncdc.noaa.gov/pub/data/gsod/", focal_year, "/gsod_", focal_year, ".tar")

  # Assemble your preferred file name once it's downloaded
  focal_file <- paste0("gsod_", focal_year, ".tar")

  # Download the data
  utils::download.file(url = focal_url, destfile = focal_file, method = "curl")
}
1
This message isn’t required but can be nice! Downloading data can take a long time so including a progress message can re-assure you that your R session hasn’t crashed
2
To create a working URL you’ll likely need to click an example data file URL and try to exactly mimic its format
3
This step again isn’t required but can let you exert a useful level of control over the naming convention of your data file(s)

In slightly more complicated contexts, you’ll need to make a request via an Application Programming Interface (“API”). As the name might suggest, these platforms serve as a bridge between some application and code. Using such a method to download data is a–relatively–frequent occurrence in synthesis work.

Here we’ll demonstrate an API call for NOAA’s Tides and Currents data.

# Load needed packages
## install.packages("librarian")
librarian::shelf(httr, jsonlite)

# Define a 'custom function' to fetch desired data
fetch_tide <- function(station_id, product = "predictions", datum = "MLLW", time_zone = "lst_ldt", units = "english", interval = "h", format = "json"){

  # Custom error flags

  # Get a few key dates (relative to today)
  yesterday <- Sys.Date() - 1
  two_days_from_now <- Sys.Date() + 2

  # Adjust begin/end dates
  begin_date <- format(yesterday, "%Y%m%d")
  end_date <- format(two_days_from_now, "%Y%m%d")
  
  # Construct the API URL
  tide_url <- paste0(
    "https://api.tidesandcurrents.noaa.gov/api/prod/datagetter?",
    "product=", product,
    "&application=NOS.COOPS.TAC.WL",
    "&begin_date=", begin_date,
    "&end_date=", end_date,
    "&datum=", datum,
    "&station=", station_id,
    "&time_zone=", time_zone,
    "&units=", units,
    "&interval=", interval,
    "&format=", format)

  # Make the API request
  response <- httr::GET(url = tide_url)
  
  # If the request is successful...
  if(httr::status_code(response) == 200){
    
    # Parse the JSON response
    tide_data <- jsonlite::fromJSON(httr::content(response, "text", encoding = "UTF-8"))

    # And return it
    return(tide_data)

  # Otherwise...
  } else {

  # Pass the error message back to the user
  stop("Failed to fetch tide data\nStatus code: ", httr::status_code(response))

  }
}

# Invoke the function
tide_df <- fetch_tide(station_id = "9411340")
1
When you do need to make an API call, a custom function is a great way of standardizing your entries. This way you only need to figure out how to do the call once and from then on you can lean on the (likely more familiar) syntax of the language in which you wrote the function!
2
We’re excluding error checks for simplicity’s sake but you will want to code informative error checks. Basically you want to consider inputs to the function that would break it and pre-emptively stop the function (with an informative message) when those malformed inputs are received
3
Just like the batch download, we need to assemble the URL that the API is expecting

While many ecologists are trained in programming languages like R or Python, some operations require the Command Line Interface (“CLI”; a.k.a. “shell”, “bash”, “terminal”, etc.). Don’t worry if you’re new to this language! There are a lot of good resources for learning the fundamentals, including The Carpentries’ workshop “The Unix Shell”.

Below we demonstrate download via command line for NASA OMI/Aura Sulfur Dioxide (SO2). The OMI science team produces this Level-3 Aura/OMI Global OMSO2e Data Products (0.25 degree Latitude/Longitude grids) for atmospheric analysis.

Step 1: Using the “subset/Get Data” tab on the right-hand side of the data page, generate a list of file names for your specified target area and time period. Download the list of links as a TXT file named “list.txt”. Be sure to document the target area and temporal coverage you selected in your data inventory table.

https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0802_v003-2023m0804t120832.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]
https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0805_v003-2023m0807t093718.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]
https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0806_v003-2023m0809t092629.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]
https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0807_v003-2023m0809t092635.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]
https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0808_v003-2023m0810t092721.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]
https://acdisc.gesdisc.eosdis.nasa.gov/opendap/HDF-EOS5/ncml/Aura_OMI_Level3/OMSO2e.003/2023/OMI-Aura_L3-OMSO2e_2023m0809_v003-2023m0811t101920.he5.ncml.nc4?ColumnAmountSO2[119:659][0:1439],lat[119:659],lon[0:1439]

Step 2: Open a command line window and execute the wget command. Replace the placeholder for username and password with your EarthData login credentials.

wget -nc --load-cookies ..\.urs_cookies --save-cookies ..\.urs_cookies --keep-session-cookies --user=XXX --password=XXX
--content-disposition -i list.txt

If you encounter any issue, follow this step-by-step guide on using wget and curl specifically with the GES DISC data system.

Activity: Data Download (~25 mins)
  • Each member work on the data that you have been assigned.
  • Discuss with your group how to collaborate on coding without creating merge conflicts
    • Many right answers here so discuss the pros/cons of each and pick one that feels best for your group!
  • Write a script for your group to download data using your chosen method
  • Zoom rooms for each download method will be available. You are encouraged to join the room that corresponds to your chosen method to discuss with others working on the same approach.
    • If no datasets in your group’s inventory need the download method you chose, try to run the example codes provided

Data Format and Structure

CSV and TXT are common formats for data storage. In addition, formats like NetCDF, HDF5, Matlab, and Rdata/RDS are frequently used in research, along with spatial datasets such as geotiff, shapefiles, and raster files (refer to the spatial module for more details).

In R, information about data’s structure can be retrieved in a variety of ways. We’ll go ahead and load the dplyr package as well to be able to include one Tidyverse option for evaluating structure.

# Load needed libraries
library(dplyr)

# Define URL as an object
dt_url <- "https://pasta.lternet.edu/package/data/eml/knb-lter-sbc/77/10/f32823fba432f58f66c06b589b7efac6"

# Read it into R
lobster_df <- read.csv(file = dt_url)

# Displays the first few rows of the dataset to provide a quick overview
head(lobster_df)

#Provides a summary of the dataset, including basic statistics for each variable
summary(lobster_df)

# Displays the structure of the dataset, including data types and a preview of the data
str(lobster_df)

# Offers a transposed, compact overview of the dataset's structure, data types, and values
dplyr::glimpse(lobster_df)

# Checks if there are any missing values (NA) in the dataset
anyNA(lobster_df)
1
This is the Tidyverse equivalent of str. The information looks different because the data are coerced into a “tibble” behind the scenes

Additional Resources

Papers & Documents

Workshops & Courses

Websites