Broward Replica Trips – Data Summary

Survey basics

The data covers 7,728,125 unlinked trips for 2,132,763 people across 987,137 households.

Demographic information

Each trip is tagged with demographic information about the traveler and their household. These are the attributes available:

• resident_type: either “core” (lives in the region) or “visitor” (does not live in the region). Demographic information is only provided for core residents! [about 93% of unlinked trips are taken by core residents]

• age: age of traveler (given as an exact integer age)

• employment: either “employed”, “unemployed”, or “not in labor force”

• race/ethnicity: if hispanic, then simply “hispanic or latino origin”; if not hispanic, race breakdown includes “white”, “black”, “Asian”, “American Indian or Alaska native”, “native Hawaiian or Pacific islander”, “two or more races”, and “other”

• household income: income of the traveler’s household, in exact integer dollars

• household vehicles: vehicles in the traveler’s household; either 0, 1, 2, or 3+

• household size: size of traveler’s household, in an exact integer people

Information is also provided on the traveler’s home and work locations. Both home and work block group are available (again, only for core residents), along with geographic coordinates for these block groups.

Trip information

Each unlinked trip, regardless of resident type, is defined by the following attributes:

• mode: trip mode; one of “private auto”, “carpool”, “public transit”, “walking”, “biking”, “commercial”, “on-demand auto”, or “other”

• travel purpose: trip purpose; one of “home”, “work”, “school”, “social”, “commercial”, “shop”, “eat”, “rereation”, “lodging”, “maintenance”, “stage”, or “other”

• previous activity type: activity at the origin, featuring the same options as travel purpose

• distance: trip distance, in meters (converted from 1/10 miles)

• duration: trip duration, in seconds (converted from whole minutes)

Trips are also tagged with times: exact times for both start and end of trip are provided, as are generalized integer hours of departure/arrival.

Specific trip origin and destination information is available, as well: trips are tagged with origin and destination block groups, as well as geographic coordinates for the block groups. If a trip originates or ends outside of the region, this information is not provided.

Finally, if applicable and available, information on the vehicle used is also available: options include “passenger car”, “medium commercial”, and “heavy commercial”. Though there are fields for fuel type and fuel technology, these are always unknown or omitted.

Discussion on relationship to MSP “Ideal Dataset”

This is not an exhaustive list, but rather initial ideas on how this dataset compares to the MSP Ideal.

Features that make this dataset “ideal”:

• Specific origin/destination information for each trip
  • Allows for an understanding of where people travel and informs route identification. Such specifics are not always available (e.g. NHTS, which only indicates a state)
• Ability to link trips
  • Allows for better definition of travel behavior; by linking trips, we can get a more accurate count of true “trips” per day.
• Demographic information for the travelers
  • Allows for more informed travel behavior modeling, if applicable; demographics can be used in modeling mode choice, travel budgets, and likelihood of taking a specific-purpose trip.

Features that make this dataset “un-ideal”:

• No path information
  • Despite knowing origin/destination block groups, we do not know the path the traveler took to reach their destination. Though we can assume the path (i.e., use the shortest path), or attempt to estimate it (i.e., compare k-shortest paths to the travel time and pick the closest match), these methods are imprecise and computationally intensive.

Other ideas are welcome!