This week, the Math/Stat Department brought New York Times Graphics Editor Amanda Cox for the penultimate lecture in its colloquia series. The talk, entitled “Data Visualization (and a Little Bit of Math) at the New York Times”, gave a brief overview the how and why behind the various diagrams, interactive animations, charts, and maps of the online and paper issues of the Times.
The fairly large 30-person graphics staff of the New York Times is composed mainly of cartographers, illustrators, and programmers. Editors, designers, and reporters meet daily to gather facts and decide how to react to the day’s news. Graphics can range from animations illustrating an event (crane collapse, plane into Hudson) to scientific data to election results to (increasingly) the economy. Cox’s work is a fair balance between data reporting and analysis and graphic design. According to Cox, staffers also “spend about a quarter of their time simply revamping submitted graphics, commonly science graphics,” highlighting an example from the College’s very own Steve Wang’s work on baseball manager statistics.
Cox noted that the the Times’ graphics staff “often tries to layer more information or add in an extra layer of what’s going on in the story [by] pulling something forward while pushing another back.” In a graph tracking top home run hitters , Cox showed how this visual gave more of a background and context to Barry Bonds’ home run streak by also overlaying other top hitters. The chart then becomes an interesting deviation from the standard (boring) table while also highlighting a few trends readers may not have noticed before (Bonds’ record doesn’t level off with age as others do).
As for the “little bit of math” in the lecture, Cox mentioned that most of the math involved in graphics reporting was “straightforward, elementary to middle school mathematics. There’s a lot you can do with sports statistics, exit polls, and simple probabilities.” Surprisingly, the Times graphics staff does not employ a professional statistician and only occasionally calls in PhDs or graduate students when higher-level mathematics is necessary. This diagram illustrating inflation and consumer spending , for instance, made use of a Voronoi diagram or a series of cells surrounding a given set of points with each cell containing area closest to its own central point.
One audience member brought up a contentious point over the use of circles in a map of Medal Counts for Olympics past. The majority of readers can’t accurately perceive magnitude differences between the medal counts since the area of a circle is not linearly related to its radius. Cox acknowledged and countered the comment by explaining that the geography dimension of the graph, especially in a graphic about the Olympics, was more important to the visualization of the data.
The Q & A session following the talk further focused on how the NYTimes gauges the efficacy of their graphics and how academics themselves can improve their own graphics in scientific publications. The talk certainly provided something new to think about as we pick up our morning papers.