The bookdown package makes creating a book within RStudio shockingly easy. However, if you’re used to writing in LaTeX, there are some humps and bumps that need to be smoothed out before working with bookdown is as easy as it purports to be. In this talk, we will walk through the knowledge bookdown assumes you have before starting, its little idiosyncrasies, how it compares to LaTeX, and how to use it to make a highly customized book (e.
Read more →
Often times, when we are ‘looking’ at data in plots, we find a set of interesting points or identify a pattern that is interesting. What is the significance of a finding like that? Visual inference gives us protocols that help us to quantify the strength of a visual finding in a framework similar to to confirmatory statistical hypothesis testing. Visual inference helps analysts determine if structure is real or spurious. In this framework, plots take on the role of test statistics, and human cognition the role of statistical tests.
Read more →
“A picture is worth 1000 words” is as true in scientific communication as it is in other domains, but why are pictures such an effective way to communicate numerical information? In this talk, I’ll give an overview of the human visual system, focusing on the areas which influence our perception of graphics. We’ll discuss the hardware and “software” of human vision, and then I’ll give a brief overview of some of the research on the perception of statistical charts and graphs.
Read more →
The Vega-Lite framework offers the traditional grammar of graphics, rendered in the browser. It also offers a first implementation (at least within Vega-Lite) of a grammar of interactive graphics. The vegawidget package (not yet available on CRAN) provides a means to create and render Vega-Lite specifications using R.
In this presentation, we will go through a tutorial overview of vegawidget. To participate, you will need only a laptop with a modern browser (such as Chrome or Firefox), and an internet connection.
Read more →
Mining temporal-context data for information is often inhibited by a multitude of time formats: irregular or multiple time intervals, multiple observational units or repeated measurements on multiple individuals, heterogeneous data types, nested and crossed factors indicating hierarchical sub-groups. Time series models, in particular, the software supporting time series forecasting makes strict assumptions on data that needs to be provided, typically a matrix of numeric data with an implicit time index. Going from raw data to model-ready data is painful.
Read more →
