intro_ndarrays.qmd

---
engine: julia
title: Basics of multidimensional data analysis
author: "Felix Cremer, Fabian Gans"
institute: "Max Planck Institute for Biogeochemistry 


[NFDI4Earth Advancing Tools](https://nfdi4earth.de/2facilitate/advancing-tools)"
format: 
    revealjs:
        logo: images/logo2_mpi_nfdi4earth.png
        number-sections: false
        footer: "[Datacubes in Julia](https://juliadatacubes.github.io/datacubes_in_julia_workshop/)"

execute:
    echo: true
---

## Data cubes

![Earth System Data Cube](images/esdl.png){alt="A three by three layout of data cubes with different data types."}

## Data formats

- Geotiff
- Zarr
- NetCDF
- ...

## Chunking

::: {#fig-chunking layout-ncol=2}

![Spatial chunking](images/img1.png){alt="Representation of the spatial chunking where every map is put together in memory."}

![Box chunking](images/img2.png){alt="Representation of a box chunking where yellow boxes indicate that both temporal and spatial neighbours are close together in memory."}

Different chunking structures
:::


## Pyramids

```{julia}
#| eval: false
using PyramidScheme
using Rasters, RasterDataSources
using ArchGDAL
rastas = Raster(CHELSA{Climate},:tas, lazy=true, month=1)
rastasscale = rastas.metadata["scale"] .* rastas .+ rastas.metadata["offset"]
pyrtas = Pyramid(rastasscale)
using GLMakie
plot(pyrtas, colormap=Reverse(:RdYlBu))
```

## Pyramid example
{{< video images/zoom.mp4 >}}