Anscombe's quartet comprises four datasets that have identical simple statistical properties, yet appear very different when graphed. The datasets make a very good case for visualizing data, as the major differences cannot be found with a simple statistical analysis.