kellytdunn / Cluster-Analysis

A cluster analysis of transit passengers responding to a survey

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Cluster Analysis

A cluster analysis of transit passengers responding to a survey

The 2021 Passenger Experience Survey (n=4890) asked transit passengers questions about their travel behavior, demographics, and satisfaction with Sound Transit. I used a cluster analysis to identify clusters of passengers who had similar travel behaviors. This analysis used the following metrics:

  • How often they ride
  • How long they have been a rider
  • Modes they ride
  • Access to a vehicle
  • Typical trip purpose
  • Times of day they ride

The analysis showed that there are four main clusters of passengers based on travel behavior:

  • Cluster 1: Peak commuters (25%). They primarily ride Link and ST Express at peak hours, and ride more frequently than average. Newer riders are disproportionately represented here. Some of these commuters do not have vehicle access.
  • Cluster 2: Weekend Warriors (5%). They ride less frequently during the week and overall but more on weekends, primarily for recreation and social outings. Some also ride at peak hours for work. They ride all modes and most have vehicle access.
  • Cluster 3: Convenience Riders (65%). They ride less frequently but have been ST riders for a long time. They primarily ride for recreation and social outings, as well as for airport trips. Most travel is on weekends and weekday afternoons and is on Link. Most have vehicle access.
  • Cluster 4: Night Owls (5%). They ride frequently, especially evenings, late nights, and weekends. They are less likely to have vehicle access and primarily ride Link and ST Express.

I then looked at demographic and satisfaction data for each cluster and found that the four clusters vary somewhat in demographics and travel behavior:

Cluster Peak Commuters Weekend Warriors Convenience Riders Night owls
Demographics Representative of general ridership, Distributed throughout ST District More men, Slightly older, More likely to have a disability Primarily in Pierce County Slightly older, Slightly higher incomes, Live in Seattle and close-in suburbs More likely to be gender-nonbinary, Younger, Lower incomes, Live primarily in Central Seattle and North King County
Satisfaction Lower scores for overall grade and many sub-metrics Lower scores for overall grade and many sub-metrics Higher overall grade Average or slightly above average scores for sub-metrics Lower scores for all metrics

About

A cluster analysis of transit passengers responding to a survey


Languages

Language:R 100.0%