Here is a simple map posted by New York City Economic Development Corporation (NYCEDC) that attempts to illustrate the geography of pizza joints in New York City. While, I love that the map almost looks like a pizza and the concept is interesting (but not uncommon), here’s how the map in weak in so many ways:
- Direction: Add North arrow. Dude, this is Map 101!
- Boroughs: Label the thick green line that represents borough boundaries, and of course, label the boroughs!
A picture is worth 1000 images. Remember that famous saying? Well this map does exactly the opposite:
- Off the top of my head: What does 0-5 mean? The obvious response is that it is the number of the places that sell pizza, but you have go digging into the map description to confirm it.
- Thinking it through: A huge technical flaw in this map is that the number of pizza places is not adjusted to the population or the area of each zip code.
So, a large zip code in Staten Island with 16 pizza places among 1000 people is represented in the same dark orange as the bustling uptown zip code in Manhattan with 16 pizza places for 60 places! It’s comparable to coloring the population of Russia in the same category as Bangladesh’s population based without accounting for density1. Do notice that the authors address this issue by saying, “land area and residential population play a big role Land area and residential population play a big role in the number of pizza places in a particular zip.
- Off the top of my head: Besides the postal services, who really uses zip codes? Would propose a happy hour in 10003 or East Village?
- Thinking it through: In addition to zip codes being impractical for the human context, from my sleepless night with vector-based projects, I know that map with zip codes are a GIS faux pas that is best avoided. It is very unusual to find GIS data such as population, housing structure, median income etc., in zip code level. These data are mostly available at block group or county levels. Here is the problem- zip code boundaries do not completely overlay with block group or county boundaries. For example, both Sparta and Ravenscroft counties in Tennessee lie in zip code 38583. So it possible for one zip code to have that 2/3 of its population from one county, and 1/3 from another. In this case, you have to interpolate the data from two separate counties before you can do further analysis. If you have ArcGIS 10, you can add the Areal Interpolation tool, but keep in mind; this is a very time-consuming process.
- Too much white space in the background. Additional information and labels mentions above ( in the “Where?” section) would help manage the while background better.
- It would be a great idea to add images of some popular pizza places from the Red zip codes (i.e. the highest numbers of pizza joints).
3 TAKE HOME MESSAGE
- Direction: Never ever forget, the compass or north arrow
- Design: Manage your labels well to give the viewer enough information to 2 key questions: Where is this place in your map? What does your map represents (units)?
- Also, keep in mind that ancillary information such as source, coordinate system and projection are also very important elements of a professional map that can be less visible, but present.Avoid, I repeat, AVOID using data at the zip code level. Most urban data (in the U.S.) is available at the county or block group level, not zip code level. So you would have to buy ArcGIS 10 and/or install the Areal Interpolation tool and go through an time-consuming ordeal of interpolating data from county to zip code or vice versa.
1. Based on information available through a quick Google search on world population ranking on Dec. 9, 2011 – the population of Russia is 142,914,136 and Bangladesh follows close at 142,319,000.