Wednesday, September 7, 2011

Reaction : Graphical Perception

Objective of this paper is to establish a scientific foundation for graphical analysis and representation of data. The paper follows a systematic approach of explaining the concept of graphical perception,which author precisely describe as "visual decoding of information encoded on graphs". The two step approach involves collecting basic visual tasks and ordering them according to human perception. Paper illustrate efficient ways of extracting info from different graphs like bar,pie, Cartesian,volume graphs etc and then derive a hypothesis that impose an ordering on elementary perceptual tasks on the basis of accuracy that brings perception closer to actual encoded values. Ordering is supported with proper reasoning, experimentation and psycho-physics. I liked the way author explains the supposition that length judgment is more accurate than area, which is further accurate than volume, with a psycho-physics theory based on judgement of position and frame of framed rectangle. Mathematical formulas and experiment data does prove the viability of position-length experiment, however i feel that author should have provided more information on subjects as human perception and understanding plays a major role.The idea of proving the validity of approach by offering replacements( dot,grouped charts, framed rectangle) for commonly-used graphs, is very intuitive and novel. Author accepts shortcomings of the theory but also leaves room for improvements. Overall its a very informative article and makes a good attempt to establish a theory which was never in focus of researchers.