Wednesday, November 30, 2011

NZ Unemployment Continues to Climb & LOESS curves

Today's data viz challenge has been using R (stats software) to build a chart of New Zealand's unemployment with a LOESS curve showing the big picture behind the fluctuating data points.  LOESS curves are clever, and R has a function for them, giving you a curving trend line in a way you just couldn't achieve with MS Excel.

Although the task in hand is charting up some data, I was surprised to see just how dramatic the unemployment story is - jobseeker numbers have almost quadrupled in the past three years!

Is this better or worse than other countries?  The mood of the nation doesn't seem as pessimistic as elsewhere, so perhaps others are even worse off.

Tuesday, November 29, 2011

Creating Step Charts in R

I'm determined to boost my programming skills - this looks like the only way to create truly customised data visualisation. So my weekend projects tend to consist of many hours tapping away at the keyboard until I get something I'm happy with.
Today I've created a chart showing the price of a postage stamp in New Zealand. Unlike most price-by-time charts, postage stamps suit step charts, because prices are fixed for long periods between step changes.
Now the chart below might not look like rocket science, but bear in mind that neither Excel nor Illustrator has a built-in step chart type, as Nathan Yau points out in his data viz book "Visualize This". So to create this in something like Excel, you would have to use some kind of work-around.
The downside is that to use the wider array of chart types available in R (free statistics software), you need to get your hands dirty with a wee bit of programming. Since that's what I'm trying to learn, it all works out quite nicely!