An interactive shiny-based application for exploring graph- constrained changepoint detection with the gfpop package.

OverviewInstallationTestingAttributions

Overview

This is a graphical user interface (GUI) for the gfpop package.

Generally, gfpop is an R package implementing an algorithm that lets users detect changepoints in their data. Unlike related packages (e.g. changepoint), gfpop works under the assumption that its users have some general idea about the patterns of their changepoints. For example, you might expect that your data alternates between “up” and “down” changepoints, each with a mean change of at least 1. If you can represent your assumptions about your changepoints as a directed graph, gfpop can take that constraint graph as an input and very efficiently predict changepoints in your data.

This GUI for gfpop lets you draw a constraint graph and edit its parameters interactively. Then, it will run the gfpop algorithm and let you interactively see changepoint differences.

This is an in-progress summer project. See the project description at Google Summer of Code 2020.

For more complete documentation, see the pkgdown site.

Installation

To install:

devtools::install_github("julianstanley/gfpopgui")

To run, you can either visit the web application, or run:

gfpopgui::run_app()

See the application for more detailed usage information. Documentation is still in progress.

Testing

Uses a combination of testing frameworks. testthat and shiny::testServer are used for most functionality and run via CI.

Other functionality is tested on the most recent version of Chrome on OSX Catalina and Windows 10 via SauceLabs.

Tests are also run on the latest releases of Windows 10, macOS, and Ubuntu 16.04 via GitHub actions.

Attributions

Project supported by the R Project for Statistical Computing With funding provided by Google Summer of Code
alt text alt text

Open Source testing license generously provided by SauceLabs.com :————————-: Testing Powered By SauceLabs