De tablas a gráficos, presentaciones en Stata

Stata has few facilities for presentation graphics, largely because it is aimed at more or less advanced technical people rather than lay or business people. However, for many purposes even very advanced users often would benefit from having access to more tools for display, especially of raw data or of results that come in tabular form with text labels attached.

In this talk I survey various programs, mostly written for Stata 6.0.

tabm, tabsort and tabsplit allow the tabulation of (respectively) multiple variables with similar scales, variables sorted by rows or columns or both, and string variables split into parts.

hplot is a graphical workhorse that can produce a variety of horizontally labelled plots for data, including W.S. Cleveland's dot charts or dot plots; variations on them with continuous rather than dotted lines; D.R. McNeil's horizontal parallel line plots; and displays for showing key quantities with or without confidence intervals.

hbar produces a horizontal bar charts. Bars for different variables are stacked with base at 0, depending on whether values are positive or negative. Optionally, bars may be horizontal lines with or without vertical line s at either or both ends.

The symbols for hplot go beyond the standard Stata set as implemented in graph, symbol() and the shadings for hbar include no shading and invisible bars, the last allowing some special effects.

One of the main advantages of hplot and hbar is that plenty of space is available for longer text labels. You are not obliged by space or by Stata rules to use labels no more than 8 characters in length. Equally simple but equally important is that all text is written horizontally, thus avoiding `giraffe graphics', in which you must hold your head at varying angles to read the text.

Given hplot and hbar it is easy to write drivers for specialised variants. tabhplot and tabhbar represent tables of frequencies, while cihplot displays confidence intervals.