GS Recurrence Project

[ml874]

Application that uncovers historical time periods similar to a user-specified query time period / similarity function, with respect to a handful of factors / time series. Example question that the application will try to answer is: what are 10 time periods most similar to 3/20/20 - 9/1/20 (2020 Tech Melt-Up) of the same length with respect to NASDAQ (+ other economic factors)?

GS_Recurrence provides an API, along with out-of-the-box visualization and analytical features.

GS_RecurrenceUI provides a GUI, which uses ipywidgets / plotly to render only in Jupyter Notebooks (depicted below).

Also provides functionality for calculating how factors move after the found similar time periods by providing a table of basic statistics such as average / standard deviation of return, p-value denoting how significant those returns are, etc.