A Python package to preprocess time series

Disclaimer: This package is WIP. Do not take any APIs for granted.

tspreprocess

Time series can contain noise, may be sampled under a non fitting rate or just need to be compressed. tspreprocess is a library for such preprocessing tasks. It contains tools to transform and clean time series data for better analyses.

In detail, we are planning to add methods to do

• Denoising
• Compression
• Resampling
• ...

Our goal is to make this the most comprehensive time series preprocessing library.

Installation

Clone the repo, cd into it and install it with pip locally

git clone https://github.com/MaxBenChrist/tspreprocess
cd tspreprocess
pip install -e .

You can run the test suite by

python setup.py test

Relation to tsfresh

This package will based on the data formats from the python feature extraction pacakge tsfresh (https://github.com/blue-yonder/tsfresh), allowing a seamless integration between both packages.