This repository contains the raw data and analysis that are the supporting information to the bletl paper.

Contents

The unprocessed raw data was placed in the data directory.

The notebooks directory contains Jupyter notebooks for all analyses and figures in the manuscript.

Figures were not git-committed, but the extracted time series features and t-SNE were exported to the results folder.

Installation

While the bletl package itself may be installed from PyPI, this repository with the full analysis also depends on PyMC3 that can be a little tricky to install. For this reason we provide an environment.yml file that can be used to reproduce a Python environment with all the dependencies:

conda env create -f environment.yml

The new environment is named bletl_env and can be activated with conda activate bletl_env.

After that a Jupyter notebook server can be launched with jupyter notebook.

References

• bletl source code
• bletl documentation