StatPlayground: Exploratory Statistics through Direct Manipulation of Visualizations
What is StatPlayground?
StatPlayground is a prototype, built as a potential solution for a lack of adequate statistical literacy. StatPlayground performs statistical analysis automatically by checking the statistical assumptions (e.g., normality, homoscedasticity) and, based on which assumptions are satisfied, choosing and then performing the appropriate significance test. StatPlayground also gives the user the ability to learn various statistical analysis concepts in an exploratory manner by 1) changing the data properties to see how it impacts other data properties and the computed inferential statistics, and 2) changing the computed inferential statistics to see how the data characteristics change.
We believe StatPlayground has the ability to improve certain statistical literacy skills: the ability to make sense of statistical information, identify the relationships between statistical concepts, data awareness, and understanding statistical procedures better.
StatPlayground is a research project by Krishna Subramanian, M. Sc., and is supervised by Prof. Dr. Jan Borchers. It is funded in part by the German B-IT foundation. Jeanine Bonot is currently doing a Master's thesis, in which she will investigate interaction design techniques for a more fine-grained control of the parameters.