Quote:
Originally Posted by Verge
Hence the inaccuracy (or arbitrary nature) of density figures-- urban area calculations determine only population above a certain density threshold. And no I dont think Atlanta is the least dense big city (I never said that)-- but it is still probably among the least. This is true of most modern cities-- Phoenix, Dallas, Houston, etc.-- Surprisingly LA is among the most dense--
|
My comment wasn't directed at you. FWIW, the issue with density isn't accuracy, which is function of our measurement tools and is also improvable, but sensitively to arbitrary thresholds (both the level of density we choose as the cutoff as well as the continuous area to define it - I'm not saying you disagree with this, but the two concepts are not interchangeable). But this is an issue with pretty much any metric as the "goodness" of a measure is dependent on agreement with the purpose it serves.
Your example of LA is instructive on this point as it can be characterized as having greater or lesser density than NY depending on where you choose to set the boundaries. LA lacks NY's peak density, but its urban/suburban density drops off at a lower rate, so aggregate density can be greater if the area of comparison is large enough. Atlanta is something like this (in terms of the distribution pattern - it is obviously not in the same league of either city in absolute terms). Atlanta lacks a region of very high sustained density, but the density drop off in its inner suburbs is also not that great. There is a pretty significant drop-off as you transition to the outer suburbs, hence the low MSA density, but the point is you could remove 80% of the metro region by county area and still have a respectable "city" of nearly 4 million. Or to put it another way: The 10 inner counties of Atlanta metro area have almost exactly the same population and area (i.e. the same aggregate density) as the Boston MSA. Again, I don't mean to suggest this isn't something you already know, since you mentioned an interest in this topic, but I want to underscore the point about how results are a function of the chosen methodology and even using a consistent standard (e.g. census) doesn't mean you don't get misleading results since you can't hold geography and other factors constant.