Editorial

Abstract

Event History Data

Vis. Methods

Vis. Tools

Comparing Pencils

Further development

Difficulties

New tools

Acknowledgements

References

Case Studies Index

Visualisation of historical events using Lexis pencils

Editorial Introduction

A very common data type in social research is the categorical in which individual cases are coded by some attribute (gender, place of residence, work status, etc.). Sometimes such a variable might be an attempt to locate the individual on some underlying scale that can be assumed continuous such as "social class", but more often than not the categories are strictly nominal and many of arithmetic operations do no apply to them.

Methods for the statistical analysis of such data have been developed for many years and are nowadays widely used, but methods for their visualization available in most packages are primitive. Typically these may consist of a bar chart or "pie" diagram (often misused) and not much else, and the difficulty of visualization is compounded if the time change in such categorical variables also needs to be visualized.

Brian Francis and John Pritchard's Case Study presents the Lexis pencil as a way of visualizing a moderate number of individual time changes in a number of categorical variables. The technique and its development is of interest for its own sake, but also by the way it illustrates three more general points of interest:

1. 1. In common with several other visualization techniques developed in this series of case studies, it builds on a well known, tried and tested graphical device, the Lexis diagram. It may well be that this types of background is an almost necessary condition for a "new" graphic to be accepted by a user community;
2. 2. Although the graphic in static form is useful, its analytical usefulness is increased immeasurably by interaction in zooming and animation. Again, this seems to be a feature of all the successful graphic devices in these case studies;
3. 3. Although the original intention was to build a syustem in an existing visualization system (AVS) in practice time and experience has shown that this proved very awkward and has led to the development of a stand alone program. Regrettably, this also seems to be a feature of almost all the other case studies, and it has the undesirable consequence that, in order for visualization methods appropriate to social science applications to be developed, considerable programming skills seem to be necessary. The implications of this fact for the further development of visualization in the social sciences should be obvious.