Shower of bullets
Apr 25, 2007
Here's one of those infographics that makes the reader work hard (via Dustin J). The graphic in its full glory is here; it's much too large to be reproduced, and I have clipped off the bottom half.
Much to the designer's credit, he extracted data of interest, rather than trying to cram everything onto the page. In particular, he was most interested in the distribution of deaths among different age groups, the types of deaths (suicides, homicides) and the identities of the deceased (race, gender).
Just like the election fraud graphic, such rich data lend themselves to multiple levels of aggregation. Here, the designer focuses on the most detailed level, making it easiest to see facts like "among the 18-25 age group, there were 6 black men murdered per day".
However, it takes much more attention to notice higher-level facts like "homicides per day are relatively flat across age groups while suicides heavily skew toward 40+".
In the junkart version, I decided to emphasize the more aggregated data, showing the number of deaths of each type across age groups. The detailed break-down of race and gender is shoved into parentheses, as they can be omitted by less serious readers.
The reader who discovers that the homicide/suicide pattern described above may surmise that homicide gunfire deaths are more "random" while suicides, being premeditated, may affect older people disproportionately. More research would be needed to confirm such and other suspicions.
Source: "An Accounting of Daily Gun Deaths", New York Times, April 21 2007.
I have to say, I really like the original chart. The graphics personalize it, draw me in, and focus on the disproportionate number of older white males committing suicide. It made me want to go analyze the data, something that few charts can do.
Posted by: Sam Adams | Apr 25, 2007 at 08:01 AM
This is the kind of data that requires more than one chart, the "multiple levels of aggregation" cited by Kaiser. The NYT graphic has tried to put everything into one graphic, and one motivation seems to be the use of bullet images as the main graphical element.
Kaiser's line chart shows the breakdown of shooting deaths by age group and classification much more effectively than the original. Unfortunately the NYT graphic has left out information, indicated as "?" in Kaiser's labels.
Posted by: Jon Peltier | Apr 25, 2007 at 08:40 AM
The way age ranges are being used seems bad to me. There are four datapoints covering different spans of time: 17 years (<17), 7 years (18 to 25), 13 years (26 to 39), and 35+ (>40 -- and that's assuming a 75 year life span). How can those ranges be compared in any sort of useful way?
Posted by: Mike | Apr 25, 2007 at 10:28 AM
Coming from a culture less familiar with bullets, my first thought was "Why a chart of dildos?"
Posted by: Doire | Apr 25, 2007 at 10:43 AM
Mike, they would have some justifying to do if queried on those exact ages, but it's not wrong in general to have periods that are of different lengths, if those periods come with mental images of "types of citizen".
It's true that the 40+ group have a whole thirty five years in which to die by gunshots, while the 18-25 group have only seven, but if the survey is addressing and disputing a perception that 18-25 year olds suffer more gun deaths, then it's a legitimate way to group them IMO.
For a quick alternative that uses the available data, perhaps Kaiser or someone else could redesign the junkart version so that it normalises each group by the estimated number of years in the group, to see how that affects the story that the graph tells?
Posted by: derek | Apr 25, 2007 at 01:56 PM
I agree with Mike. There is a strong correlation between the number of people in an age group and the number of people who get shot in that age group.
In a line chart which emphasizes comparisons between ages, all other significant variables should be kept constant between the groups. Ignoring the distribution of the underlying data items is misleading in this case.
The murder rate is much higher in the 18-25 group, but neither visualization shows this directly.
Posted by: bryan | Apr 25, 2007 at 04:40 PM
Thanks, Doire. I saw the same thing, but I'm glad that I'm not the only one!
Posted by: IB a Math Teacher | Apr 25, 2007 at 09:13 PM
Good observation on the baseline. Gives me an opportunity to rant against "per day" statistics, like "X people die from Y disease every day". (Not everyone is at risk of disease Y, and not every day is the same.) As pointed out here, the relevant basis is the distribution of population by age, not the fact that there are 365 days in a year.
Posted by: Kaiser | Apr 25, 2007 at 11:49 PM
I think both yours and the NYT graphics would benefit from a separate table or graph of the totals (numbers below guessed)
Posted by: Timz | May 01, 2007 at 02:34 PM