DataCube Exploration and Visualization

Hello!
I am trying to retrieve Sentinel-2 data filtering by date and location. I need to retrieve 3 images before a certain date and 3 images after a certain date of a specific location (of a landslide event). I load the collection (as in the attached code snippet), and I wish to see how many images were loaded in the DataCube and their acquisition date.

Additional information about the acquired imagery would be also very helpful for my case, as I’m creating a training dataset with before and after images of landslide events.

I wanted to ask whether it’s possible to explore the metadata of a loaded DataCube, and if not, are there plans to extend the capabilities to be able to do so?

Additionally I found this great tool to visualize DataCubes: LexCube - Leipzig Explorer of Earth Data Cubes. I believe it would be a very useful addition for data exploration - would you consider adding such a tool to the Platform?

For similar topic: a similar question was asked by @sulova.andrea but the reply didn’t really cover my question.

Thanks in advance for your thoughts, ideas and help regarding this issue!

Best,
Julia :slight_smile: :smiling_face:

Hi Julia,
I guess that the question you are referring from @sulova.andrea is this one right?

Currently we do no have a direct way to get the metadata only in openEO Platform as far as I know. Please correct me if I am wrong @stefaan.lippens @sean.hoyal @daniel.thiex .

Anyway, the simplest workaround I could propose you to use is to select a really small area in your AOI and the temporal extent covering your range of interest and save the result as a netCDF . In this way you will get the information about the available dates, let me know if you need a sample code for this.

@m.mohr maybe the inspect process could be used to return the metadata of the cube without the actual data? I remember we discussed about this functionality time ago.

It would likely (depends on the implementation) return the metadata and an excerpt of the data, but that should be fine, right?

If the use case is getting access to full metadata, then a catalog search that is similar to the load_collection call would give you exactly this.
I’m not sure if all of our collections consistently link to the STAC/OpenSearch catalogs behind them, but that’s perhaps something we could work on.
Note that we also add ‘derived-from’ links to STAC metadata of generated results, so you can also go back to original metadata for inputs.