Chris Knox
Data Editor and Head of Data Journalism
NZ Herald
Chris Knox is a data journalist at the New Zealand Herald. R is his primary tool for seeking out and understanding datasets that might be the basis of a story. Trained as an atmospheric chemist, Chris made the jump to scientific visualisation early in his career and has spent the past twenty years working in the fields of data visualisation and more recently journalism. His role as a data journalist is to help the public connect with diverse datasets in engaging ways.
How R powers data journalism at the New Zealand Herald
Data analysis within the fast-paced environment of the newsroom usually involves one of two types of analysis:
- Pulling together often messy data from multiple sources as quickly as possible for use in a single story. This can involve multiple analysis iterations with other journalists and domain experts before deciding what the key story is.
- Rapidly processing and analysing data after a scheduled release e.g GDP data, the Budget, election results. Usually, but not always, the structure of the data can be inferred from previous releases. This means that theoretically the code to process, analyse and publish from the new data can be prepared in advance and, if all goes well, a data piece can be published a few minutes after the data is released.
R’s combination of support for easy interactive exploration combined with excellent reproducible workflows means that it hits the sweet spot for both of these analysis scenarios.
In this talk, I will outline how I use R at the NZ Herald to efficiently and reproducibly work with data ranging from election results, census data and NZ gazette records to olympic rowers’ alma mater and the price of cheese. I rely heavily on quarto and the targets pipeline tool and will walk through a couple of cases in which these have been used to process, analyse, and publish the data for a couple of recent NZ Herald articles.