Some data journalists are providing replication files with their stories-- enabling readers to download a package of data and scripts that reproduce a story's findings. I will show how to adapt these techniques to the reporting stage of a story--a practice I'm calling “preplication.” On a recent project, I supplied open-source preplication packages to a dozen Wall Street firms I was investigating, to allow them to critique the statistical scripts I wrote to rank their trading practices. This computerized evidence “confrontation” helped me clean up errors in my scripts and ensured the story held fewer surprises for them...or me. These preplication files then became the replication files we posted for our readers. In this session, we'll examine and run these files in Rstudio. You'll learn about pitfalls and best practices in such pre-publication file-sharing, which are matters of journalism practice not confined to a particular application or language. While we're doing this class in R, it won't require specific knowledge of R or any other language.
Bill Alpert is a Senior Editor at Barron’s, The Dow Jones Business and Financial Weekly, where since 1984 he has published nerdy stories and investigative projects of use to investors. His exposes have also been followed by stock fraud prosecutions, government recoveries under the False Claims Act, and reforms in such diverse areas as “back-door” stock exchange listings, for-profit colleges, and the Magnitsky Act that sanctions international human rights violators.
No tipsheets have yet been uploaded for this event.