Data visualisation
One of the important aspects of the work of an actuary is not just the analysis, but to get the point across.
Data Visualisation
This website gives examples of the best (and worst) ways to use graphics for statistics
Here is a blog on data visualisation. And here's another called junkcharts which was recommended in science on 13 October.
Here is a RSS feed to a good blog by Stephen Few. He focuses on Business Intelligence, but it is quite interesting. Here is a link to his page on the Business Intelligence Network. There is an open Data Visualisation 2006 competition which has just finished. In his articles he's reviewing the winning entries until early September.
Online Resources
This document gives some other examples using R and gives the code for creating them (not checked as I am not an R expert!)
This gallary of plots using R may be rather slow but contains many good ideas
This page shows how to create custom colours within R in order to customise your graphics.
For those of you interested in open source software, here is a brief overview of data visualization tools.
Examples
 3D Surface Density Plots
3D charts are usually a little dodgy for displaying data, but they do look pretty. The "rgl" package allows the user to rotate, zoom, etc. on the graph with mouse movements. Therefore, this makes visualising data in more than one dimension much easier for the analyst.
I used this type of plot to visualise the effect of a complex reinsurance contract involving several variables (i.e. not just claims costs). It was useful to check whether or not the calculations were correct.
Here is a function to plot this type of graph.

 AMO And Hurricanes Graph
This graph appeared in a recent series of Lloyd's papers.
See this page for the code to create this graph.

 Funnel Charts
These are useful for illustrating uncertainty in future projections  in particular as a graphical output from monte carlo simulations. Here is a function to plot this type of graph.

 Scatterplot with Marginal Histograms
A useful plot for looking at the relationship between two variables, for example when considering copulas.
Here is a function to plot this type of graph.
It is straightforward to turn the scatterplot part of this graph into a contour plot of densities.

Comments (0)
You don't have permission to comment on this page.