Development and Evaluation of Interactive Smartphone App for Visualization of Bayesian Inference Using R-Shiny

Abstract

Purpose

This work studies the porting of B.T.V.T.I., a Bayesian inference visualization program for PCs, to smartphones and tablets. The function of the program’s operations was evaluated after porting and a user survey was conducted to evaluate its utility in learning and understanding Bayesian inference.

Method

The R and Shiny programming languages were used for the porting, while the survey was carried out via Google Forms.

Result

The BTVTI-R program, with its user interface re-arranged for smartphones, afforded a nearly identical reproduction of B.T.V.T.I. operations. However, screen size restrictions disabled the porting of those B.T.V.T.I. techniques that had several types of input data. A survey question asked to ascertain how interesting this app had made Bayesian statistics for its users revealed that 65.6% of respondents were interested in it.

Discussion

The connectivity of BTVTI-R with the server, believed to be unstable when the program was left unattended for a short period of time, was a cause of concern as it could impair the convenience of operations. As indicated by the survey results from respondents familiar with Bayesian inference, this application could, in combination with books and classroom lectures, be a useful tool for learning this topic. Further, it is possible that this simple application, if integrated into a smartphone, might spark interest about Bayesian inference in the minds of novice learners.