I also attached the extension - with the sample app and data.
Please mark the appropriate replies as helpful / correct so our team and other members know that your question(s) has been answered to your satisfaction.
The boxplot is a great way to visualize the shape of data sets using statistical methods to find and depict range, median, outliers, and quartile-based variability/whiskers. Here's a good explanation of it from Wikipedia.
The way the data is brought in is only a tiny bit unusual. We want all of the numeric values with NO AGGREGATION so you actually want to add 2 dimensions - first dimension is the actual dimension (in the example provided it's a Quarter, so Q1, Q2, etc). Second dimension is the set of numeric values you want the chart to crunch on, in the example provided it's Revenue.
In order to be able to replicate exactly what Bostock did and see they match, I am ingesting his sample CSV in the app provided; However, because his csv is pivoted, I used a crosstable function in the loadscript to unpivot and get the data in a way we normally consume it.