We have been busy working on several projects so I have not had a chance to blog recently. Since it is GIS Day, however, I think I owe another blog post given the proliferation of maps on the web these days, especially in the mainstream media. There are now many data journalists that have been empowered by numerous free and easy to use tools to complement their analyses with a map. Does that mean you should always include a map though? When I look at a map now I try to think about whether the map provides real value to the analysis or whether it is simply there because they could put it there.

I recently spotted an article from Vox that I was referred to through the Yahoo home page that depicted the “ten deadliest interstates in America.” Being in the traffic safety field, I am always interested to check out stories analyzing collision data. However, I quickly found myself questioning the effectiveness of the article, especially the use of maps. The author states the purpose of the work was to determine where fatal collisions are most likely to occur and “whether there are certain stretches of highway that seem to have a disproportionate number of collisions given their size.” She ends up determining that in 2013 Interstate 285, the Atlanta, Georgia outer beltway is the deadliest highway stretch in the nation with ’26 accidents, 29 fatals and 3.5 accidents per 10 miles.’ Perhaps a useful observation, but I had a bevy of critiques regarding the analysis. As I started listing issues, however, this blog post started to become a long-winded discussion of those flaws and diverted from my original reason for writing. Therefore I will spare you all the critiques, but I imagine for many people the article will cause confusion and they will lose interest well before reaching the end of the article if they are not in the traffic safety field. (I even conducted an independent survey of 1 person that confirmed my suspicions!)

My initial thought though was about the map at the top of the article displaying all the highway collisions in 2013. What does a general reader gain from viewing that map? As I looked at the map, it reminded me of this xkcd comic I have seen reference multiple times in other blogs:

Comic interpretation of many heat maps.

The joke being that despite the content of the map many end up just looking like a population density map. I felt the main map from the Vox article is another form of this pet peeve #208, where the collision map simply becomes the equivalent of a map of interstate highways with ‘population’ clusters around the major cities. Showing all those collisions without any context does not give any perspective for ‘deadly’ highways and is really just a dot representation of the roads.

This issue is certainly not just limited to this article though as even I have been guilty of providing a national map of fatal collisions in the same density form like the comic. I believe I did so more out of the belief that people would think it looks interesting rather than believing it provided any real value. I am sure the author of the Vox article probably felt the same, but now when I see maps I try to take a step back and think critically of whether they are actually informative. Are general readers doing the same? Probably not. Maybe I am wrong and having the map does give many readers perspective as to the number of collisions that are occurring on the interstates. I am just not sure that’s the case.

In the end, I think the most important thing to keep in mind is that each of those dots on the map are actually terrible events where one or more persons lost their lives too early. Despite any deficiencies with the article, if it can help raise more awareness to the situation it will probably at least have a small positive influence on the long road to eliminating traffic deaths.

Just keep that comic in mind next time you look at a similar map though. It might just give you a new perspective.

 

Co-Founder of RoadSafe GIS. Bringing collision data, GIS and cloud technology together. Formerly of UC Berkeley SafeTREC and ESRI.