CPI-DaliPie by Keith Ng
What is driving inflation? Over the past few years, there's been a lot of media coverage of rising prices. This is an attempt to put CPI changes into context by intuitively showing the relationship between the CPI weighting basket and the changes of each group. This visualisation allows users to zoom in to every part of the CPI, to see how individual product groups have changed over time, and how they affect the higher-level groups. The "Dalipie" is a variation of the Nightingale chart. Unlike the Nightingale chart, however, both variables are dependent variables. The second variable maps to the area of each slice (if the second variable increase by 50%, the area of that slice increases by 50% - this is not the same as the radius increasing by 50%!). It's called DaliPie because it's salient effect is disproportionality. The fact that the slices do not line up jar with user expectations of what a pie graph looks like, and this jarring immediately draws attention to the graph's main feature. An element's total area and its disproportionality are immediately apparent: e.g. "This element is big/small, but it has grown disproportionately." The disproportionality is paired with animated disaggregation. You can see how each slice breaks down into more volatile slices (or conversely, when volatile slices are joined together, they average out). A lot of time has been spent on the animation because that implicity explains, purely visually, how a CPI actually works - a weighted average of changes across a broad group of items. It was a very conscious choice to use a pie-like graph. When you zoom in to a small category, even though it's filling up half the screen, the curve of it's shape tells you how big that group is in relation to the whole circle. Top-level context is retained in a way that is not possible with any other shape.