Title: Visualising and exploring models interactively with Shiny
Authors: Catherine Hurley - Maynooth University (Ireland) [presenting]
Abstract: Two new interactive tools for model exploration are discussed. Both are implemented as Shiny-based R packages ERSA and condvis which are available on CRAN. ERSA is aimed at the student and teacher of linear regression models. It uses linked displays of model summary tables and drill-down parallel coordinate plots of case-based information for families of models. With a point and click interface, students explore the effects of dropping predictors and cases on model fits and their visualisations. The second package, condvis, provides a way of examining model behaviour for fits ranging from linear models, to Bayesian fits to black-box models. It displays fits on sections of data space, together with observed data near the section. Sections are either chosen interactively by the user, or calculated to reveal some feature of the model. Such visualisation offers a way of examining model behaviour, checking model stability and juxtaposing multiple model fits, leading to improved understanding and possibly discovery of improved fits.