Skip to content

Latest commit

 

History

History
49 lines (33 loc) · 3 KB

raster.md

File metadata and controls

49 lines (33 loc) · 3 KB
Raw

Contributing Raster Data

Data structure

All raster data for the site is stored as Cloud Optimized GeoTIFF (COG). One way to validate that data is in the proper format is using rio-cogeo:

  • First, check that it passes validation with rio cogeo validate my_raster.tif
  • Then ensure that it has a nodata value set and that it matches the value which represents non-valid pixels within your GeoTIFF. You can see the nodata value like so:
rio info my_raster.tif --nodata

note: nan values in the data will not be treated as non-valid pixels unless the nodata tag is nan.

This same library can also create a Cloud Optimized GeoTIFF with the following command:

rio cogeo create my_raster.tif my_cog_raster.tif

Naming convention

New raster files are added to the dashboard manually so the naming convention is rather liberal. The only requirement is that for date-specific data, the file name must include the date, formatted as:

  • YYYYMM for monthly data
  • YYYY_MM_DD for sub-monthly data (anything with daily or semi-periodic data)

If the file doesn't have global coverage, please use a portion of the file name to indicate the spotlight area it covers. We provide data for the following spotlight areas. For example:

my_tif_tk_2020_02_19.tif # Data for Tokyo on February 19th, 2020.

Metadata

When sending the data, please include the following information

  • A short description (1-2 sentences) of the data to be included on the dashboard.
  • The time and spatial domain covered by the dataset
  • Best practices/standards of color maps for visualizing the data

Delivery mechanism

There are three mechanisms for making raster data available through this API:

  • publicly available: any publicly available COG can be accessed through this API. Because the API is run on Amazon Web Services in the us-east-1 region, data hosted on S3 in this region will have faster response times to the API.
  • send to API maintainers: if you'd like to keep the source data stored privately, please contact [email protected] or [email protected], and we can discuss other hosting options for the data.
  • upload directly: some science partners have direct S3 upload access. Those partners can upload to s3://covid-eo-data/[dataset_folder] where [dataset_folder] is an S3 folder containing the data. Each dataset should have a 1-1 relationship with a folder.

Visualization

Once ingested or otherwise made accessible, the data is available as map tiles as detailed in the API documentation. There are a variety of parameters that can be used to customize the visualization, in particular, a number of colormaps. The remaining parameter descriptions are shown here.