Morrizzzzz
commented
2 months ago Now that I worked on updating Ep 5 and 6 I noticed that for the accessing data part you actually need to have quite some python knowledge (classes, methods, dictionaries etc.).
Also I find the STAC documentation not always very easy to follow (It took me some time to get the idea of ItemCollection and Item) and it is very specific. We train people doing something in a specific platform which for a different platform would be something completely different.
An idea that would be very drastic, but which would make a lot of sense now that I thought of it multiple times (I get these ideas while cycling btw).
The participants start our workshop with the expectation that they will learn something about how to work with raster and vector data in python. Now a significant part at the beginning of the workshop is dedicated accessing satellite images from a specific platform, whereas for some this might not be very interesting (they just want to work with raster images, which can be anything they already have (aerial photograph, Corine landcover, elevation models etc. right?).
Obviously, working with satellite images still makes a lot of sense, but the whole part of obtaining the data might be better place later in the course or even optional. I would propose that for the raster part we just prepare a dataset (like we did for the vector episode) and work on that one.
proposed order:
ep1 to ep4 would remain the same.
ep7 vector data
ep6 raster data intro
ep8 crop raster data
ep9 raster calculations
ep10 zonal stats (why don't we integrate this with ep9?)
ep11 parallel raster computations
ep 5 Access satellite imagery using Python
@fnattino @rogerkuou What do you think?
Looking at the material I think it makes a lot of sense to switch episode 6 and 7. This way we can use the vector data generated in episode 7 to manipulate the rasters in episode 6. The danger is obviously that this messes up things.