Introduction

wavy is a library for time series analysis in Python. It is a wrapper around the pandas package.

The goal is to provide a simple and intuitive interface for time series analysis, while providing a high level of flexibility and extensibility.

The processing starts with a DataFrame with the data to be analyzed, where the index is the time and the columns are the different variables.

When creating the Panel object, the user needs to specify three things:

  • lookback - the number of time steps to look back

  • gap - the number of time steps to skip

  • horizon - the number of time steps to look ahead

../_images/panel.png

For example, lets suppose that we have a DataFrame with 12 time indexes and we want to create a Panel with lookback=5, gap=1 and horizon=2.

../_images/panel_divided.png

At the end we get a Panel for x and y with 3 frames each.