search

EO-College / cubes-and-clouds

This is the official repository for the online course 'Cubes & Clouds'
Creative Commons Attribution 4.0 International
32 stars 12 forks source link

Improvement exercise 2.3 reduce #33

Closed przell closed 8 months ago

przell commented 8 months ago

@clausmichele, In the lecture about reducing: https://github.com/EO-College/cubes-and-clouds/blob/main/lectures/2.3_data_access/exercises/23_data_access_reduce.ipynb

clausmichele commented 8 months ago
  1. Plotting is easy: image
  2. There seems to be some +-inf derived from NaNs getting into the NDVI formula
  3. If we want the number, it means that it will download the whole set of data. Doable, but I'd rather reduce the time range or the spatial AOI to reduce the downloaded data
  4. Same as before. Better to reduce the AOI and increase the time range?
przell commented 8 months ago
  1. Plotting is fine!
  2. There have also been some other values like 2.43 for example
  3. Can we not reduce all the time steps to one number, the mean, and then plot that as a time series. The download should then only be one value per time step? ...
  4. ... then just plot this time series.
clausmichele commented 8 months ago
  1. I will investigate
  2. To aggregate the values you firstly need them! And since we're computing locally, it would anyway have to download data first.
przell commented 8 months ago
  1. Yeah, forgot about the local computation part. Either we reduce the AOI drastically, or we just don't show a time series. Up to you ;)
clausmichele commented 8 months ago

Solved in the last commit!