Last year, a new version of dplyr was released. It handles non-standard evaluation using a new framework: tidyeval, developed by Hadley Wickham and Lionel Henry. Since then, tidyeval has been incorporated also into tidyr and ggplot2. Tidyeval will be useful if you write code that uses dplyr on data-frames with column names that you don’t know until runtime. This might be the case if you are writing package functions – this also might be the case if you are writing a shiny app that works on user-uploaded data-frames.
Read more →
Census data provide an important snapshot of information about a country at different times throughout its history. The evolving and adapting nature of census questions creates a complicated data management problem, and this problem is exaggerated when the goal is to answer questions across several years simultaneously. However, recent developments in methods for data analysis and visualization provide a window for developing a more clean interface for census data exploration. I will present a web-based Shiny application that allows users to browse U.
Read more →
Almost everybody working in R is working with R packages. The package system in R is a convenient way to bundle your R functions and data and make it available to the wider community. In recent years the support for package development has increased. We will be discussing functionality of RStudio and the devtools package to support R development. We will work on a small R package together as an example.
Read more →
As consultants on a wide variety of projects across many majors, a common oversight we encounter is a failure to examine the residuals. This is particularly true when the client is performing the analysis in R. We were inspired by the residual panel in SAS to create an R package that easily provides users with a similar panel of plots. The ggResidpanel package in R is intended to give a single view of diagnostic plots for checking the key underlying assumptions of linear models.
Read more →
Graphics are very effective for communicating numerical information quickly and efficiently, but many of the design choices we make are based on subjective measures, such as personal taste or conventions of the discipline rather than objective criteria. We’ll talk briefly about perceptual principles such as preattentive features and gestalt heuristics, and then discuss the design and results of a factorial experiment examining the effect of plot aesthetics such as color and trend lines on participants’ assessment of ambiguous data displays.
Read more →
