Skip to content
ELLIIIT Dataset Docs ELLIIIT Dataset Docs ELLIIIT Dataset Docs

Post-Processing Workspace

The repository includes a user workspace at post-processing/.

Use that folder for your own analysis after the processed datasets are available in results/.

Typical content in post-processing/:

  • exploratory notebooks built on top of the RF and acoustic datasets
  • custom scripts for feature extraction, filtering, or plotting
  • figures and tables for reports, papers, or presentations
  • experiment comparison notebooks and follow-up analysis

The goal is to keep:

  • published or downloaded datasets in results/
  • shared parsing and tutorial code in processing/
  • your own follow-up work in post-processing/

Minimal Getting Started

Section titled “Minimal Getting Started”

Clone the repository:

Terminal window
git clone https://github.com/techtile-by-dramco/ELLIIIT-dataset-26.git
cd ELLIIIT-dataset-26

Create a Python environment and install basic notebook dependencies:

Terminal window
python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install jupyterlab matplotlib numpy requests xarray pyyaml h5netcdf

Download the published acoustic datasets into results/:

Terminal window
python processing/dataset-download/download_acoustic_datasets.py --list
python processing/dataset-download/download_acoustic_datasets.py --experiment-id EXP003

Make sure the processed RF dataset is also present in results/, for example:

results/csi_<experiment_id>.nc

Then start Jupyter:

Terminal window
jupyter lab

Useful tutorial notebooks to copy from or use as reference:

  • processing/tutorials/tutorial_xarray_structure.ipynb
  • processing/tutorials/tutorial_acoustic_xarray_structure.ipynb
  • processing/tutorials/tutorial_rf_acoustic_position.ipynb

Suggested Folder Usage

Section titled “Suggested Folder Usage”

Create your own files under post-processing/, for example:

post-processing/my-analysis.ipynb
post-processing/compare_experiments.py
post-processing/figures/

Recommended workflow:

  • keep the shared tutorial notebooks unchanged unless you are improving the shared examples
  • keep personal or experiment-specific analysis in post-processing/
  • keep the processed .nc inputs in results/

Contributing Back

Section titled “Contributing Back”

If your post-processing scripts, notebooks, or figures are useful for other users of the dataset, create a branch and open a pull request against the main repository.

That is the preferred path when:

  • the notebook or script is reusable
  • the figures improve the shared docs
  • the analysis clarifies the interpretation of the dataset
  • the work fixes a bug in an existing workflow

Wait until the results are in good shape, then open the pull request so the work can be reviewed and merged.

The repository copy of these instructions is also available in post-processing/README.md.