Overview

Cities can earn a maximum ParkScore of 100.

For easy comparison and at-a-glance assessment, each city is also given a rating of one to five park benches. One bench means the park system needs major improvement, while five benches means the park system is outstanding.

In evaluating park systems, experts at The Trust for Public Land considered land owned by regional, state, and federal agencies within the 75 most populous U.S. cities—including school playgrounds open to the public and greenways that function as parks.

Our analysis is based on three important characteristics of an effective park system: acreage, facilities and investment, and access.

Acreage

ParkScore awards each city points for acreage based on two equally weighted measures: median park size and park acres as a percentage of city area. Factoring park acreage into each city’s ParkScore helps account for the importance of large “destination parks” that serve many users who live farther than ten minutes’ walking distance.

Median park size is calculated using park inventories acquired from park-owning agencies within each city. In our national sample, median park size ranges from 0.6 acres to 16.8 acres, with a median of 5.0 acres.

Park acres as a percentage of adjusted city area is calculated using data collected in an annual survey conducted by The Trust for Public Land’s Center for City Park Excellence. We removed unpopulated railyard and airport areas from the baseline city land area. In our national sample, park acres as a percent of city area ranges from 2.0 percent to 86.0 percent, with a median of 9.1 percent.

Facilities and investment

ParkScore awards each city points for facilities and investment based on two equally weighted measures: total spending per resident and an average of per-capita provision of four key facilities - basketball hoops, dog parks, playgrounds, and recreation and senior centers. This data is collected in an annual survey conducted by the Center for City Park Excellence.

Facilities were chosen because of the breadth of users served as well as the ease of accurate counting of these measures. In our national sample, basketball hoops per 10,000 residents ranges from 0.7 to 7.5, with a median of 2.3; dog parks per 100,000 residents ranges from 0 to 5.4 with a median of 0.7; playgrounds per 10,000 residents ranges from 0.8 to 5.1 with a median of 2.3; and recreation and senior centers per 20,000 residents ranges from 0.1 to 2.5 with a median of 0.7.

Spending per resident is calculated from a three-year average (FY 2010, FY 2012, and FY 2013/2014, depending on a city’s fiscal calendar) to minimize the effect of annual fluctuations. Spending figures include capital and operational spending by all agencies that own parkland within the city limits, including federal, state, and county agencies. In our national sample, spending per resident ranges from $19 to $343, with a median of $83.

Access

ParkScore awards each city points for access based on the percentage of the population living within a ten-minute (half-mile) walk of a public park. The half-mile is defined as entirely within the public road network and uninterrupted by physical barriers such as highways, train tracks, rivers, and fences.
Learn more about the methodology behind the ten-minute walk to a public park

In our national sample, the percentage of the population living within a ten-minute walk of a public park ranges from 27 percent to 99 percent, with a median of 71 percent.

Scoring

The scoring system recognizes the accomplishments of cities that have made significant investments in their parks without holding dissimilar cities to an unrealistic standard. It enables detailed analysis and allows cities to increase their ParkScore through incremental improvements to different aspects of their park systems.

To determine a city’s ParkScore, we assigned points in three categories: acreage, services and investment, and access.

  • Acreage: 20 points for median park size, and 20 points for park acres as a percentage of city area
  • Facilities and investment: 20 points for spending per resident and 20 points for the average of the four key facilities per-capita (basketball hoops, dog parks, playgrounds, and recreation and senior centers).
  • Access: 40 points for percentage of the population living within a ten-minute walk of a public park

Points for each statistic are assigned by breaking the data range established by our national sample into 20 brackets, with the lowest bracket receiving the least points and the highest bracket receiving the most points.

Each city’s total points—out of a maximum of 120—are then normalized to a ParkScore of up to 100.

Outliers

To prevent outliers from skewing the results, the top bracket for each measure includes all values equal to more than double the median of the data range. For example, spending per resident in our 75-city national sample ranges from $19 to $343, with a median of $83. Most cities spend between $50 and $150 per resident. To control distortion from local anomalies, all cities that spend more than double the median value ($166 per person) are assigned to the highest bracket and receive 20 points.

With the top bracket thus defined, the parameters for the remaining brackets are established so that each bracket comprises an equal portion of the remaining data range.

This protocol applies to all categories except access, which has no outliers.

Mapping

To map access to parks and open space, ParkScore first identifies gaps in park availability, then determines which gaps represent the most urgent need for parkland.

Access gaps are based on a service area representing a ten-minute walk (see “Access” above to learn more). To map park need, we combined three differently weighted demographic profiles:

  • Population density - 50%
  • Density of children age 19 and younger - 25%
  • Density of individuals in households with income less than 75% of city median income - 25%

Each city’s park need is mapped from data collected in the 2014 Forecast Census block groups provided by Esri. Learn more.

You can find the interactive ParkScore maps on the Explore page.

You can also explore Historical ParkScore Rankings.

The CDC created a model that predicts block group-level small-area estimates of childhood obesity for children based upon the age, gender, and racial/ethnic composition, and local community contexts of each block group. These results were published in the journal Preventing Chronic Disease, Volume 10 - May 02, 2013. This is the first such study to examine childhood obesity prevalence nationwide and The Trust for Public Land will be showcasing the data through The Trust for Public Land ParkScore® index. These estimates provide valuable data to identify priority areas for feasible local intervention goals such as providing access to parks for kids to recreate and exercise. The data can be found on the explore page in the map viewer. It should be noted that these obesity estimates do not take into account park access and thus no correlation can be made between park access and the obesity estimates presented here.