pytsal

An easy to use low-code open-source python framework for Time Series analysis, visualization and forecasting along with AutoTS.

Why was pytsal created?

I was deeply inspired by pycaret which is an amazing library for Machine Learning, and I wanted to create a similar library for Time Series Analysis.

Therefore, the interface and features provided are very similar to pycaret but focused and customized towards Time Series.

What does pytsal mean?

Pytsal is the abbreviation for Python Time Series Analysis Library

Overview

Image source

Features

Checklist of features the library currently offers and plans to offer.

Convention used below: Feature [status]

• Time series data loaders [partial]
• Time series preprocessing [partial]
• Time series modelling
  • Forecasting
    • Holt Winter [completed]
    • ARIMA [in progress]
    • Facebook Prophet [planned]
  • Classification [planned]
  • Anomaly Detection
    • Brutlag [completed]
• Time series visualization [v1 completed]
• Time series validation [v1 completed]
• AutoTS
  • Forecasting [v1 completed]

Getting Started

The following instructions will get you a copy of the project and ready for use for your python projects.

Installation

Quick Access

• Download from PyPi.org

  pip install pytsal
  

Developer Style

• Requires Python version >=3.6

• Clone this repository using the command:

  git clone https://github.com/KrishnanSG/pytsal.git
  cd pytsal
  

• Then install the library using the command:

  python setup.py install
  

Examples & Tutorials

Tutorials on how to use the library can be found under the examples folder

The tutorials clearly explain how to use the library and also provide basic guide to understand time series analysis.

• Forecasting tutorial
• Anomaly detection tutorial

Stability

The library isn't mature or stable for production use yet.

The best use of the library currently would be for non production use and rapid prototyping.

Current Contributors

Made with contributors-img.

Contribution

Contributions are always welcomed, it would be great to have people use and contribute to this project so as to help users understand and benefit from the library.

How to contribute

• Create an issue: If you have a new feature in mind, feel free to open an issue and add a short description on what that feature could be.
• Create a PR: If you have a bug fix, enhancement or new feature addition, create a Pull Request and the maintainers of the repo, would review and merge them.

What can be contributed?

• Datasets
• Source code enhancement
• Documentation

Author

• Krishnan S G - @KrishnanSG