2 Talks: Michael Kane on fc package and Kyle Burkett on R at Progressive

NOTE the change to 5/2

Talk 1:
Generalized function composition and pipe construction with the fc package*

The "pipe" operator is a convenient way to define sequences of functions. Implementations like magrittr's %>% work by modifying unevaluated expressions to create functions that are called serially using the convention that intermediate values are stored as a variable named "." This approach results in "pipelines" that may be easier to read (but are often slower) when compared to "unpiped" alternatives. In this talk, I'll show an alternative pipe implementation that does not rely on non-standard evaluation. Pipeline construction with this new package is slightly less compact than tidyverse equivalents but often yields more performant functions whose definitions are easy to read.

Michael Kane is an Assistant Professor in Yale University's Biostatistics Department. His research interests are in the areas of scalable statistical/machine learning, statistical computing, and early-phase clinical trials.

Talk 2:
How do you deploy and manage R at a fortune 200 company? Do you use Windows or Linux? What version? What packages? What difficulties do you face and how do you make RAM available with large datasets? Join us as Kyle Burkett, IT Systems Analyst Lead, talks a bit about his experiences as the owner and maintainer of all things 'R' at Progressive Insurance.