TrendPy: Time Series Regression with Python

[Zoufiné Lauer-Baré]

Dear SymPy Community,

 I am pleased to present a next open source Python Project of Interactive Calculus that uses SymPy.

https://github.com/zolabar/trendPy

TrendPy is a Python package tailored to easily perform powerful regressions with linear, polynomial and sinusoidal ansatzes. Further a three-parameter "free" regression with user ansatz is possible. The free regression part is based on SymPy!

The TrendPy packages is object oriented features classical aspects as returning the coefficients, coefficient of determination of the training data.

Further, a web application is provided, which allows to apply different regression models on data uploaded by the user, it extends some features of Excel's trendline.

https://trendpy.herokuapp.com/

or

https://mybinder.org/v2/gh/zolabar/trendPy/HEAD?urlpath=voila%2Frender%2F/trendpy_webapp.ipynb

Test-Data in

https://github.com/zolabar/trendPy/tree/main/data

This project is based on NumPy, SciPy Jupyter Notebook, Plotly and SymPy.

It is yours to discover! Feedback is always welcome.

The project is uploaded to Zenodo and can be cited as

Lauer-Bare, Z., & Wirth, M. (2022). trendPy -  Time Series Regression with Python (Version v1.0.1) [Computer software]. https://doi.org/10.5281/zenodo.7009281

Best Wishes,

Zoufiné