How the Local Access Scores are Calculated

Local Access scores are calculated using travel demand software that uses input data on population and destinations to estimate the number of trips households are likely to make in a given day, the likely destinations of those trips, and the most direct routes connecting households to their destinations. The four steps of the model are described briefly below. For more detailed information, see the Technical Report.

  1. Trip Generation: How many trips of each type begin and end in each block? A set of equations converts the analysis zone data into a number of trips produced by and attracted to each block for each trip type. Trip production and attraction rates account for household size, number of school-age children, size of establishment (store, school), and many other factors. Rates were based on the 2012 Massachusetts Travel Survey (MTS). The Analysis Zone inputs for each trip type are described in table 3 below.
  2. Trip Distribution: How many trips go from each block to each other block? Another set of equations describes, for each analysis zone, how many trips produced by that zone end in every other analysis zone included in the study area. The “destination choice” model used to create these estimates was again based on the 2012 MTS and its information on how far residents are likely to travel for various purposes. Model input variables are on-road distance and WalkScore ® [1]. Two zones that are closer together and have higher WalkScores are more likely to have a higher exchange of trips.
  3. Mode Choice: How many of these trips might be made by walking or biking? Shorter trips are more likely to be made by walking and biking. The model accounts for this by reducing the trip probabilities as a function of distance, with longer trips less likely to occur, and thus contributing less to a given segment’s utility score. Separate walking and biking rates by trip distance were drawn from the 2012 MTS, with probabilities for walking and biking to school based on MAPC survey data on school commutes.
  4. Route Assignment: What is the most direct route for each trip? Routes were assigned based on shortest network distance between the origin and destination census blocks. The network includes all surface roadways, regardless of whether or not they currently have a sidewalk or bike facility. Limited access highways are excluded from the network, as well as specific segments where a pedestrian connection is categorically infeasible or undesirable. The trips assigned to each segment were summed up for each trip type and mode to produce eight trip/mode specific Local Access score.

These eight scores were then normalized and combined to create a composite score for each mode and for both modes together. The raw scores were rescaled to a range of 0 to 100, weighted as per Table 3, and then combined into composite scores by mode. Each of these mode composite scores was also rescaled from 0 to 100, weighted (Walk: 10, Bike: 5), and averaged to produce a Raw Composite Utility Scores. Finally, this score was rescaled to 0 to 100 to create the Composite Local Access Score.

Table 3

Field

Trip Production

Trip Attraction

Weight

Walk to School

Population 5-17 years

Public School Enrollment

10

Walk to Shopping

Households by household size

Restaurant Employees, Retail Employees

7

Walk to Parks

Households by household size

Open Space Acres

5

Walk to Transit

Households by household size

Transit Frequency

5

Bike to School

Population 5-17 years

Public School Enrollment

10

Bike to Shopping

Households by household size

Restaurant Employees, Retail Employees

7

Bike to Parks

Households by household size

Open Space Acres

5

Bike to Transit

Households by household size

Transit Frequency

5

[1] 10 most walkable neighborhoods in Boston