TRIAL-AND-ERROR APPLICATIONS FOR INTERACTIVE DATA REPRESENTATION

Abstract

This paper examines the advantages available to planners and experimental theorists of using computers to visualise their theories by generating a multiplicity of graphs interactively. The paper focuses on statistical applications, simulation and differential equation analysis. Examples show how interactive graphing systems can ease the use of complex models by immediately displaying the consequences of trial-and-error assumptions, and by revealing useful interrelationships not predicted by the theory. The paper concludes that applications exist for computer generated graphs in areas where previously hand-drawn graphs have not been practical, such as simulation, regression, estimation and forecasting.