Human vs Computer: When Visualising Data, Who Wins?
Free Public Lecture
See more events from
Can computers relieve data analysts of the arduous task of graphically diagnosing models?
Computer vision has come a long way in recent years. The models produced can now be used to automatically inspect the quality of items emerging along production lines, identify objects in photos and even navigate an autonomous vehicle.
Despite the fact that visualisation plays a major role in data analysis, the use and interpretation of graphics by data scientists/statisticians is subjective. Analysts rely almost entirely on their own judgement, years of experience and an implicit calculation of uncertainty when interpreting graphics. Considering data plots as a type of statistic allows data analysts to move away from subjectivity towards an inferential approach to reading data plots. By formalising data visualisation in this way, we explore the possibility of training a computer vision model to do this visual inference task.
In this talk, Professor Cook will give an introduction to these ideas and then present the results of computer vision model for evaluating residual plots (a type of plot used frequently in statistical analyses), comparing them to human evaluations of the same plots. Who do you think wins?