Add initial Marimo EDA for LWF datasets

[nachiketnsc]

Summary

This PR adds the initial Marimo-based exploratory data analysis (EDA) for LWF datasets.

The current work focuses on:

1. LI-COR LAI measurements
2. Individual foliage dry-weight measurements

1. LI-COR LAI dataset

LAI_Licor_3_rings_all_years.xlsx

The dataset contains Leaf Area Index measurements obtained using the LI-COR LAI-2000 system. The analysis explores measurements over time and provides interactive filtering and visualization.

The Marimo notebook includes interactive filtering by plot, subplot, plot type, and season. Missing observations and long periods without measurements are also distinguished in the temporal visualization.

2. Individual foliage dry-weight dataset

lwf_foliage_dw100_i_2026-07-30.csv

This dataset contains individual-tree foliage literfall measurements, including survey date, plot/site, species, sample/tree ID, leaf type, leaf age class, and gew100.
gew100 represents the dry weight of 100 leaves/needles.
The Marimo notebook provides interactive filtering by site, species, leaf type, and leaf age class, together with exploratory visualization of foliage measurements over time.

Status

This is an initial EDA implementation. The Marimo notebooks are still under development, with further work focused on understanding the reamining structure.

[gmgivi]

Great work @nachiketnsc! Thanks a lot for this.

[gmermoud]

@nachiketnsc how does that differ from #61, which we just merged?

[nachiketnsc]

@nachiketnsc how does that differ from #61, which we just merged?

@gmermoud The existing lfw_eda.py provides a generic inspection framework for LWF files, whereas my PR implements dataset-specific interactive EDA for the LAI and Foliage dry weight datasets. I explicitly use the known structure of each dataset to build relevant filters, process the variables appropriately, and visualize their behaviour and missing observations vs time. This is just an initial implementation; more work needs to be done for the other datasets.

[gmermoud]

@nachiketnsc how does that differ from #61, which we just merged?

@gmermoud The existing lfw_eda.py provides a generic inspection framework for LWF files, whereas my PR implements dataset-specific interactive EDA for the LAI and Foliage dry weight datasets. I explicitly use the known structure of each dataset to build relevant filters, process the variables appropriately, and visualize their behaviour and missing observations vs time. This is just an initial implementation; more work needs to be done for the other datasets.

OK, but you then confirm that you are building upon #61? @nachiketnsc