As practitioners start to build evidence and improve the presentation of their data and outcomes, questions can arise around best practices and the useful presentation of information to stakeholders in various roles.
While there are obvious points where the data can only be presented in one way, situations where there are several choices to present data can pose a very specific challenge. Practitioners must take into account what the data says about the study they are conducting, and how it will be received by its intended audience. Will stakeholders be able to follow? Will certain data points need further explanation?
There will also come a time when the data collected doesn't fit into conventional tools, and must be presented differently.
In these situations, practitioners often hear "rules" for data visualization, like "use line-plots for time-series data" or "use bar charts for counts", but when should you buck convention and break these rules?
This graphic style selection guide for mapping different types of data, by Stephen Few of Perceptual Edge, briefly provides a more nuanced look at these decisions.
View the graph selection matrix here.