The Center for Neighborhood Technology (CNT), in conjunction with the Brookings Institution’s Urban Markets Initiative, has developed an interactive mapping website that shows variables affecting household transportation costs for 52 metropolitan areas (including Columbus) at the US Census block group level of detail.
The purpose is to show people that their most affordable place to live might not be the same as the place with the most affordable housing costs. Transportation costs must be considered too. The web site shows average transportation costs for the block group based on linear regression models, and may not accurately reflect your unique situation, but it’s a really interesting starting point…and loads of fun.
For example, this is a map of housing costs as a percentage of household income. While households in Upper Arlington, Bexley, Powell, Worthington, and New Albany earn more than the median income on average, they also pay more for housing as a percentage of their income.
Below is a map of transportation costs as a percent of household income. The cost is pretty flat throughout most of the region, but on average, people downtown spend less of their income on transportation.
The following map is the combination of housing and transportation costs:
There are more good maps showing transit ridership,
automobile ownership per household,
transportation cost in dollars per month, and more.
Legends show the meaning of each color on the actual website.
The methodology is described here for the technically inclined. Basically, one of six linear regression models is used to calculate block level transportation costs for each region based on five independent neighborhood variables and three household variables. The independent variables used to predict transportation cots are as follows:
- Residential Density (households per acre)
- Average Block Size (acres)
- Transit Connectivity Index
- Employment Density
- Average Journey to Work Time
- Household Income
- Household Size
- Workers per Household
The regression model used depends on the region’s “cluster.” Regions were clustered with other similar regions based on automobile ownership and residential density. Columbus falls into the largest cluster, with fifteen other cities such as Memphis, Houston, St. Louis, Phoenix, Sacramento, and Las Vegas.
These regions have average auto ownership, low transit use, low residential density, and lower household income. They also tend to be regions with slightly higher home ownership, and they are smaller in population, with a lower fraction of people living in urban areas.
While each metropolitan area is somewhat unique at the regional level, they all show similar patterns at the neighborhood, census tract, or block group level. Primarily, it’s what we already know. Increased density leads to lower automobile ownership per household and lower transportation costs. These lower transportation costs can offset higher housing costs common in many central cities. In some places, like Columbus, you might actually find lower housing costs and lower transportation costs in many central neighborhoods.