Keywords: statistical graphics, exploratory data analysis, grammar of graphics
Interactive graphics have first been introduced by J.W. Tukey as part of explorative data analysis in the early 1970s. In the almost 50 years since then we have seen a lot of systems for interactive graphics come and go, often times involving a certain deja-vue experience. Interactive graphics advanced tremendously in the 90's and early 2000's, however the systems built during this time were usually dependent on specific graphics tool kits and operating systems. More recently, we have seen an explosion in the number of tools for interactive web graphics from R (e.g., htmlwidgets, shiny). This is great for analysts since they can leverage the best of both worlds without any knowledge of web technologies. However, these advances come at various costs that may affect the usability of these tools for exploratory and expository analysis. We will explore at least 5 different types of costs that an interface for creating interactive graphics from R may suffer from and discuss we can do to minimize those costs.