Brazil Data Cube (BDC)

[Rogerio-floresjr]

Hello everyone,

I am currently pursuing my PhD at Brazil's National Institute for Space Research (INPE) and the Research Institute for Development (IRD). My thesis aims to develop tools/services based on the BDC Datacubes (Landsat, Sentinel, CBERS) and other datasets from INPE like Climate, Deforestation, Fires, etc. The goal is to provide Earth Observation (EO) derived products on a comprehensible scale for Public Health Research, primarily focusing on vector-related information (disease occurrences by city, neighborhood, etc).

While exploring the work being done in the community, I came across the stackstac project, and I it's amazing. First and foremost, I would like to thank you for sharing this project and the time invested in it.

I have managed to use stackstac with the BDC STAC. However, I had to make some modifications to the code due to the absence of the _proj:_ extension and the BDC reference system being a custom CRS that cannot be translated to EPSG coding.

Currently, I am working on translating the custom _bdc:_ extension to better utilize the BDC tiling system between collections and to facilitate the management of other STAC datasets over the same area.

I have a few questions:

• How does stackstac handle different tiles of the same Collection? I read issue Question on tile merging logic #108 and understood that if necessary, the package should add the tiles as an extra dimension (band, time, tile, x, y). Is that correct?

• Do you think it is worthy to use stackstac with different Collections, such as a combination of Sentinel and Landsat? or also be used to incorporate other raster information like climate data? My idea here is to facilitate the extraction of varied information for a given pixel.

• Is the project still considering the implementation of geospatial routines in the long term? I am particularly interested in bringing vector-related operations to Dask using geopandas in stackstac .

I am eager to receive any suggestions.

Thank you!