Visualising uncertainty with R

I came across an article the other day by Martin Davies, explaining how important it is to visualise risk and uncertainty. A couple of weeks ago I wrote about visualising cause and effect. Martin emphasises the value of visualing uncertainty as a probability density curve, which I’ve also discussed before, and he uses the programming language R, a favourite among statisticians and pirates alike.

Martin takes multiple data points from two investments and initially summarises them with common metrics such as the mean, the maximum, etc. This gives some information, but much more is available if we plot the data itself.

Then he moves on to deriving a probability density curve for each. This is a great way to get a feel for the general shape of the uncertainty and understand how things might fare in the future. With R being a statistical language, the code for doing this is incredibly straightforward, too.

In my previous blog posts about using probability density curves I’ve suggested estimating, or speculating on, the shape of an uncertainty, which is necessary for new ventures. Martin has taken the approach of building on historical data. It’s really great to have a worked example, particularly one using readily-available tooling.

It’s also good read part of Martin’s motivation for this:

[…] simply stating risk as a number, whether that number is derived in a coherent manner or not, is […] quite professionally lacking and needs attention.

Photo by Ryan Sharp