Background

With a grant from the Federal Highway Administration’s Value Pricing Program, King County Metro Transit assembled information about multi-family residential parking use at more than 200 developments in King County over the winter and spring of 2012. Parking utilization was recorded on Tuesdays, Wednesdays, and Thursdays between midnight and 5 a.m. in all residential spaces identified by property managers in each multi-family development studied. Most of these spaces were on-site, although some property managers identified additional off-site parking areas provided for residents.

The Model

Metro then developed a statistical model to estimate parking use based on building and environmental characteristics. The model was built using regression analysis. The dependent variable was observed vehicles per occupied residential unit (or parking/unit ratio) gathered from the field data. Independent variables include:

Taken together, these variables form a strong model for predicting the parking/unit. A resulting R-square value of 0.81 indicates that 81 percent of the variation observed in parking use can be explained through these seven variables.

For more information on the data collection and modeling, see Technical Memo.