Friday, March 23, 2012

Transit Oriented Housing

If you haven't heard an aspiring (or sitting) elected official in DC hype Transit Oriented Development (TOD), then you haven't been paying enough attention. And you certainly have not spent much time around city planners, for whom this buzz word is gospel.

DC is growing, and for the most part, that has people excited. The TOD philosophy professes that the majority of the new homes, workplaces, and amenities that are built to accommodate this growth should be located within walking distance of a metro station. Whether the DC of the future has 700,000 residents or 1.5 million, taking public transportation will be an increasingly attractive option if for no other reason than that the convenience of driving will decline; roads will be more congested and parking will be more expensive (unless prices are artificially suppressed at the expense of everything else). In other words, we won't all fit unless we travel more efficiently.

A companion term to TOD, which gets just about as much play from public administrators trying to sound savvy, is Mixed-Use. Land uses that generate lots of human activity (think retail, housing, offices) should be clustered together so that you don't have to waste unnecessary time (or emit pollutants) traveling between your daily activities, and so that there is a critical mass of activity to support infrastructure investments, whether public (metro) or private (restaurants and grocery stores).

If you live within walking distance of your job, then lucky you. But you probably don't. The next best thing (sorry cyclists) is to live near a metro station and to work near a different metro station, so you can get between the two using only one mode of transportation. Most of DC's jobs are located close to metro stations, but too much of its housing is not. The market and public policy have started to recognize this problem and a big percentage of new housing is being built close to transit stations and more densely concentrated than the norm (see Columbia Heights).

DC's population density is 9,856.5 people per square mile. From the perspective of mixed-use TOD advocates, there is really no excuse for a Metro station area to have a population density lower than the city average. And yet, there are 28 stations in DC or on the border that are below 9,856.5 residents per square mile. The worst offenders (Smithsonian, Federal Triangle, the Farraguts, L'Enfant, Metro Center...) are all in the Central Business District where office demand outbids residential for the limited space available, explaining the low population density, but the stations in the next-lowest tier (Fort Totten, Southern Avenue, Brookland, Anacostia) are located in predominantly residential neighborhoods. See the full list of stations below:

To maximize the utility of the Metro system, much of DC's future growth must occur around these underutilized stations (assuming land use restrictions on the city core aren't loosened). Over the weekend I am going to begin to look at how neighborhoods around the Anacostia Metro station might change over the coming years and decades as the surrounding population density increases. In the meantime, I wanted to share this data with anyone who might find it interesting.


  1. Very good point and nice chart. I'm not familiar with the history of the Metro system, but perhaps this was also caused by the original design? The lack of development around Fort Totten always baffled me since it is a transfer station for the red and green lines, but I'm glad the Walmart and Art Center are proceeding. Hopefully that leads to much more.

  2. I think the statistic is BOGUS as it limits the view of population to within only 1/4 mile of the metro station. That should be at least 1/2 mile or actually within 1 mile. Further, it should be tempered with a view of the TYPE of development adjacent to the metro station. Does anyone think we should take down office buildings in the dense area around Farragut North and Farragut West? That is the implication because they are WAY WAY below the "resident density level" average for the region. In fact, "resident density" only measures where people SLEEP. Not where they work, shop, etc. Should these later functions be moved FURTHER away from Metro to allow from some artifical goal of having more be sleeping adjacent to metro stations? This is some planning metric run amok...

    1. I chose a 1/4 mile buffer because most people say that that's the most a typical person is willing walk on each end of a train commute. The people who live or work within that circle are the most direct beneficiaries of the transit station, and in order to spread the cost of the infrastructure investment to as many people as possible, it makes sense to build densely to maximize the number of workers/residents who will use it. You are right, that it would be interesting to see how density changes if you use a 1/2 or 1 mile buffer.

      I agree with you that the measure is misleading for office-dense areas like downtown, and I attempted to acknowledge this by saying:

      "The worst offenders (Smithsonian, Federal Triangle, the Farraguts, L'Enfant, Metro Center...) are all in the Central Business District where office demand outbids residential for the limited space available, explaining the low population density."

      The counterparts to residents are jobs and extra-curricular attractions (shopping, parks, museums, etc), and this measure does not consider the either of those latter measures of activity. I titled the post Transit Oriented Housing rather than Transit Oriented Development, recognizing that I would only be capturing one variable involved in development.

      That said, there are some metro stations on this map that you know are in predominately residential neighborhoods (Fort Totten, Southern Avenue) and the metric is an indicator of how "underdeveloped" (obviously I am using this term subjectively, with a biased set of values about how land and infrastructure should be used most efficiently) the areas around those stations are.

    2. It would be interesting to add some other factors to the analysis. Perhaps combining employment density with residential density to get some measure of the overall intensity of use.

      It might also be interesting to factor in station ridership and see how well that correlates with the surrounding land use intensity...

      Likewise, it might be interesting to compare the residential densities against an adjusted average density for DC, perhaps taking out the rivers and large parks from the equation (in order to get a more useful density number for land that's developed and developable).

      Cool map, however. The general point about housing stands.

  3. Actually, if you want vibrant neighborhoods, maybe you want a good deal of office buildings, retail, and nonresidential construction on top of adjacent to, and very near (within 1/4 mile) of MOST metro stops. People use metro when it is convenient, safe, and is efficient (as in trip chaining). When there is "just" dense residential development around metro stations, it is an advantage for the people right on top of the metro station, but it will tend to not "drag in" others that see vibrancy around that station and opportunities to walk not just to make the commute but also due other chores or meet other needs. I live in American University Park. The tenleytown station could benefit from some more dense development that included residential. But i think commercial, retail, and other development might be even more important in terms of driving actual use of the metro for commuting and other purposes. As for Fort Totten, i've looked at the neighborhood demographics and as the satellite picture clearly demonstrates, there is not much unoccupied or vacant developable land in the area. There is suburb like lower density development in place. It reflects the settlement and development patthers of 30-50 years -- well before Metro became a major influence in that neighborhood. It will take 30-50 years for "re-development" to change the neighborhood. To expect more may just be unrealistically impatient.

  4. On the bright side, Brookland is densifying near the metro station incredibly quickly. A couple projects are already open(ing): Chancellor's Row ( and the Artspace Lofts ( And one much bigger development has already begun digging its requisite pre-construction pit: the Monroe Street Market ( Across the tracks on my side of Brookland, near the 12th street corridor, is the so-called Colonel Brooks project, after the tavern being demolished to make way for it ( Of course, much of this new development has faced vociferous opposition from neighbors, but I've been heartened to see other Brooklanders express support for bringing new residents/shoppers/neighbors into the area. Hopefully the redevelopment here will be successful enough to spur similar transit-oriented development around other metro stations (I know Fort Totten has at least one big development opening soon as well, and I've been watching the Rhode Island Row apartments get closer to completion near the RI Ave. metro with every passing day). Thanks for the interesting map, and good job drawing attention to the problem; I just wanted to share some of the good news, too.