User interfaces for map-based statistical applications

Abstract

There are many statistical applications that involve users interacting with maps. R package Shiny is a new tool that facilitates the creation of a variety of interactive web applications, particularly using R graphics. Shiny is useful because it requires no prior knowledge of web development tools HTML, CSS or Javascript. I explore different R graphics packages, particularly with a view to generating and interacting with maps. I then demonstrate how a statistician might implement spatially interactive applications as well as some common pitfalls they may encounter. The TrapEditor Graphical User Interface is designed to interface with the CatchIT project and allows point-and-click data cleaning of spatial coordinates for users who may have little experience in using web applications. The Variance Estimator Graphical User Interface uses many of the same tools but also serves as an introduction to the striplet method of variance estimation for spatial surveys as well as the code required by spatial calculations. Many spatial calculations are simple in principle but difficult to implement using R. Finally I discuss some of the motivation behind the striplet method as well as an example that demonstrates that the striplet method performs much better than current variance estimators for spatial surveys.