share-a-raster

template repository for sharing a COG via github pages

This can be a convenient way to share a collection of COGs with someone for troubleshooting.

1. Click the green Use This Template button
2. Enable GitHub Pages for your repository
3. Add COGs to the cogs/ folder

Example

We use this Landsat COG from MS Planetary Computer as an example: https://landsateuwest.blob.core.windows.net/landsat-c2/level-2/standard/oli-tirs/2021/046/027/LC08_L2SP_046027_20210725_20210803_02_T1/LC08_L2SP_046027_20210725_20210803_02_T1_SR_B5.TIF

It's copied into cogs/, so it's accessible publicly without authentication: gdalinfo https://scottyhq.github.io/share-a-raster/LC08_L2SP_046027_20210725_20210803_02_T1_SR_B5.TIF

Or you can point somebody to a tiler endpoint, using developmentseed's fantastic TiTiler!

https://titiler.xyz/cog/viewer?url=https://scottyhq.github.io/share-a-raster/LC08_L2SP_046027_20210725_20210803_02_T1_SR_B5.TIF

Or open with rioxarray

import xarray as xr
URL = 'https://scottyhq.github.io/share-a-raster/LC08_faux.tif'
da = xr.open_dataset(URL, engine='rasterio')
da

Propriety data

Consider sharing a file that is obscured rather than setting up complicated authentication! https://github.com/cogeotiff/rio-faux

rio faux LC08_L2SP_046027_20210725_20210803_02_T1_SR_B5.TIF LC08_faux.tif

Limitations

GitHub isn't really designed for sharing data, so this is best for small COGs (<100MB) to share among colleagues. If you want long-term storage that is high performance look into options like cloud object storage (AWS S3 etc).

Check out https://github.com/scottyhq/zarrdata for a similar example with Zarr instead of COG

Or use a service list Felt https://felt.com/blog/raster-imagery-and-geotiffs-in-felt