|Speaker||Peter Sasieni, Imperial Cancer Research Fund, London|
Stata’s low-level graphics command gph offers tremendous flexibility to produce graphs for special purposes. Two examples are the graphics for seasonal data presented by Nick Cox at last years meeting and forest plots for meta analysis included in two recent STB submissions. As with any new programming language, writing a complete program from scratch can be daunting, and it is best to start by adapting someone elses efforts. An additional difficulty in using gph is that all Stata's normal text output is suppressed so that error messages are lost and debugging becomes problematic.
In this talk, I shall outline how I set about adapting an existing ado file to produce the graphs that I wanted for a particular application. The graph produced by the new program will, if necessary, be automatically scaled so that it can be printed in magnified form without the need for special paper.