Ohio's 3 Cs + Indy + Pitt - 2020 Urban Area Weighted Densities
 

I took a look at the urban areas of Cincinnati, Cleveland, Columbus, Indianapolis, and Pittsburgh. I went all the way down to the block level to produce a weighted density for each. Here's the formula I used: https://papers.ssrn.com/sol3/papers....act_id=3119965

Urban Area... 2020 Population... Land Area. (sqmi).. Density... Weighted Density

Cincinnati..... 1,686,744.............. 752.....…................... 2,242..... 6,379
Cleveland...... 1,712,178…........... 714………................... 2,398..... 7,341
Columbus..... 1,567,254.............. 516…........................ 3,036..... 10,253
Indianapolis.. 1,699,881.............. 722.......................... 2,354...... 6,530
Pittsburgh..... 1,745,039.............. 907…....................... 1,924...... 7,810

Are you sure you ran these numbers correctly? All of those numbers are higher than the linked article has San Fran’s UA weighted population density in 2010 (5,476).

Also, standard density does suck, sure, but weighted population density and housing unit weighted density are not the same thing. The linked article discusses the latter, not the former.

Questionable numbers, and even if the calculations were done correctly, they reveal very little about what the on the ground feel is. In no universe does Columbus feel like a denser city than Pittsburgh, Cleveland, or Cincinnati. It, by far, feels the smallest and most suburban of that grouping of cities. It's aided by the fact that it's pancake flat and can have a more uniform density. That, and the OSU student housing gives it a couple of very dense tracts. But that obviously isn't reflective of the average Columbus residential density. Weighted density can be thought of as the 'feels like' density number, and there's no way Columbus city feels like it has 10,000 ppsm, let alone the urbanized area.

This post feels like Columbus boosterism with questionable methodology.

This is because the method uses weighted housing unit density, and the census IIRC counts rentable rooms (college style private housing included) as a separate unit even if there is a common area. Columbus also has a stronger AirBnB market than the others, which will both definitely inflate Columbus’ number more vis-a-vis weighted population relative to the other cities’ on this list.

I should have been more clear. I measured population-weighted density using the formula in that paper. I don't agree with his conclusion of using housing unit-weighted density. By using the smallest measurement area i.e. census blocks I alleviated nearly every concern he brought up with using population-weighted density. Sure, housing unit-weighted density will give you an idea of the built environment, but if no one's living there you're not going to feel it. I argue concentrations of population is still the most perceptible metric, and by going all the way down to the block level you get the most accurate weighted densities possible.

Sorry for the confusion.

And what measurements do you have that Columbus isn't that dense? The calculations were done properly. Going all the way to the block level it's as accurate as possible for a population-weighted density. Sorry you can't "feel" that Columbus is 200 square miles smaller than Cleveland in its urbanized area. That's 28% smaller, and yet only 8% less populated. It's easy to see how Columbus is the most dense.

That's the whole point of population-weighted density; "undevelopable" areas don't negatively skew a larger area. A couple of very dense tracts isn't going to skew an entire urban area the other way either. This was done at the block level any way. The methodology is commonly accepted, and as far as I know has never gone down to the block level for accuracy. It's time consuming and why I only did Columbus and its regional peers.

I can't seem to find my Cincinnati file, but it was similar to Columbus and Cleveland in number of blocks:

Cleveland: 20,281 blocks
Columbus: 21,854 blocks
Indianapolis: 24,217 blocks
Pittsburgh: 36,298 blocks

At this point it's safe to say that several miles of High St. from downtown north to OSU has become the preeminent urban street in Ohio. In the last five years it's really gotten big-time, and probably smokes any comparable corridor in any sunbelt city, except maybe Miami Beach.

But outside of this area, Columbus's legacy layout doesn't lend itself to density numbers that can compare to the nearby older/bigger cities.

I actually did weighted density for the blocks bounded by the Olentangy River, the Glen Echo Ravine, the CSX railroad, and I-670 i.e. roughly the near north High Street corridor. It's 4.6 square miles with a population of 70,772; a density of 15,377 and a population-weighted density of 55,423. It only continues to densify.

As far as the rest of the urban area, I think Columbus squeezes large apartment complexes in every nook and cranny. Yes, most of it is auto-centric, but somehow that hasn't affected the traffic yet as it's still nearly non-existent.

There are also many large, in tact neighborhoods of single family homes close to or in the core. German Village, Merion Village, Olde Towne East, Clintonville, and the urban suburbs of Bexley, Grandview Heights, and Upper Arlington. The bones for Linden, Southern Orhards, and Franklinton still exist for them to be among the aforementioned neighborhoods as well.

Columbus has a remarkable consistency across the entire urbanized area. So unless you've explored the entire 516 square miles it may be hard to believe it's the most dense of these cities.

This reminds me of the talking point about Los Angeles being a denser metro than NYC, and in fact being the densest metro in the US. While it might be true, it doesn't tell much about perceived density in each city. LA has developed nearly every inch of flat land in the metro, and has done so at a pretty high level of density. Newer suburban areas have SFHs squeezed in right on top of each other, even 30-40 miles out from Downtown Los Angeles. NYC, on the other hand, has insane (for N. American standards) density in the city and some inner suburbs, but has a fair amount of lower density sprawl around the edges. So, sure, LA has the denser metro of the two, but what does that really reveal other than the uniformity of development patterns across suburbia?

In this case, Columbus is the LA of the Rust Belt. Its flat geography and relative lack of industrial areas allows for a fairly uniform level of density across the metro area. It does have some impressive density along High Street, and it's commendable to see how far that corridor between Downtown and OSU has come. But from a built form perspective, Columbus just doesn't have the same feeling of urbanity and density that you get from Cleveland, Cincinnati, and Pittsburgh. In this sense, it doesn't really matter that there are a bunch of auto-oriented apartment complexes scattered around the metro, because few people will experience those areas, and even if they do, it's not a classically urban form that would read as dense or urban. That's my opinion, at least. I appreciate the effort you've taken here, Aderwent, and I apologize for suggesting you might have erred in your calculations.

It is an anomaly that nobody saw coming. Even as recently as 20 years ago, you would have been laughed at for predicting that the Columbus MSA would zoom past the Cincinnati and Cleveland MSAs by 2030. I remember visiting The Short North in the early 2000s (2002-2005) and it was a borderline ghost town that only came alive for the monthly art gallery walks.

Both Cincinnati and Cleveland are hubs of larger regions. Columbus is more isolated, without any nearby sidekicks. Akron and Dayton are much larger than anything close to Columbus.

Considering that Columbus has basically doubled in size since the 1970s, and the fact that it's over 200 sq miles of city proper, (generally the densest part of UAs) while Cleveland, Cincinnati, and Pittsburgh have lost population with the Ohio cities around 80 sq mi and Pittsburgh at less than 60 sq mi, it's not surprising.

But also, since the 1970s, very few journalists seem to be capable of discerning the differences between MSA and legacy city populations.