The New York Times is known for its good use of data visualization. It is also known for its bad uses as well ( see my blog post from September 8, 2009 ).

The May 7, 2010 New York Times article, “British Parties Jockey to Form Governing Alliance,” by John F. Burns had an impressive use of data visualization. The following chart utilized Stephen Few’s Bullet Graph technique to describe Parliament election results.

Stephen Few developed the bullet graph specifically for use in business dashboards as a means to solve the data visualization problems caused by gauges and meters. Since that time, his original specification and use case has been expanded. His website maintains the bullet graph specification document. He last updated his specification in March 2010. In its simplest form, a bullet chart is a bar chart inside another bar chart. This binomial comparison works very well for visually analyzing the results of the British parliamentary election results. At a glance, each bullet  graph easily compares the new parliament vs the outgoing parliament seats in four categories. It is easy to see the labour party lost seats, the conservatives gained seats and that the liberal democrats and other parties basically maintained their seats. The graph also has a vertical line representing the number of seats needed to gain a majority. This is another part of the bullet graph specification used for encoding a comparative measure. In this case, the point where a majority is achieved. With this added element it is easy to see the labour party lost its majority and that the conservatives came very close to getting a majority.

Conclusion: The use of a dashboard graph type by the New York Times added simple clarity to the data in the underlying story. This is an example of data visualization done right.

For further information I strongly urge you to read the full bullet graph specification. The image below is from the latest version with all components identified.




@dmgerbino