Peninsula MobilityThe Traffic Annoyance Effect
And Other Predictors for Shuttle Ridership
Team
Katherine PhanComputer Science
John ZhaoCivil / Environmental
Engineering
Shawn LeeMaterials Science
Management Science and Engineering
Motivation
● Midterm presentation, expressed interest in the overlap between traffic data and bus/shuttle
○ Tackling congestion on local roads
● Team refined focus even more: ○ Focusing on commute trips
○ SOV trips <10 mi of big employer
TrafficShuttle /
Bus
Research Question
“How to reduce single occupancy vehicle rates for commutes of <10 miles from an employer?”
CalTrain Station
Employment Center
10 mile
s10
mile
s
Highway
Methodology
Palo AltoStanford
Bus traffic information
Shuttle traffic information
Arrival and departure analysis
Methodology (Analysis)
Marguerite Shuttle Data
Palo Alto Crosstown Shuttle Data
Arrival times per stop (time and date)
Routes and stops
Ridership (boarding) information
Relationship between traffic and shuttle ridership
Methodology (Analysis)
Marguerite Shuttle Data
Palo Alto Crosstown Shuttle Data
Real-time traffic through stops and
routes
Time estimates inside and outside
traffic
Sample Output
Palo Alto Crosstown shuttle stop locations
Palo Alto Crosstown shuttle arrival and
departure times
Travel time (without traffic) Correlation
between traffic delay and shuttle
ridership
Palo Alto Crosstown shuttle boarding
count per stop
Travel time (with traffic)
Results - Visualization Sketch
Route Visualization
Traffic along route
Passengers per stop
Results - Travel Time (minutes)
RouteScheduled Shuttle
Travel Time
Total Car Travel Time
No Traffic Non-Peak Traffic Peak Traffic
Marguerite AE-F 19.00 12.47 13.73 14.04
Marguerite U 9.00 5.23 6.13 6.51
Crosstown 50 32.77 36.28 38.77
Results - Potential Correlation
No significant correlation or with only peak hour points (bottom two graphs)
Results - Potential Significance
Slope P Value
All pointsDelay Time -0.0068 0.87
Percent Delay -0.0040 0.80
Rush hour onlyDelay Time 0.0020 0.97
Percent Delay 0.0079 0.73
No significant correlation, but correlation may improve with more data
What does this mean?
No significant correlation yet
between traffic delay and shuttle
ridership
This pattern holds with all routes
and just peak hour routesWe will have a better sense of the
relationship with more peak hour
ridership data
Deficit of Disaggregated Municipal Shuttle Ridership data
Importance of Intra Regional Commuting(OnTheMap)
Explore home-to-stop and connectivity analysis
OnTheMap: Redwood City 10.6%
OnTheMap: Mountain View 13.2%
OnTheMap: Stanford 16%
OnTheMap: Menlo Park 17.3%
OnTheMap: Palo Alto 19.3%
Future Work
1. More granular data from all partners
2. More predictors
a. Distance from stop to work
b. Distance from home to stop
c. Connections (also connectivity analysis)
d. Inconsistency of travel time
e. Frequency of shuttles
3. Primary surveys on transit systems (IRB)
4. Make a case for potential policy interventions
Thank you!Questions?
Managers Mobility Partnership
● Agreement to collaborate on transportation issues regionally
Redwood City
Palo Alto
Mountain View
Stanford
Menlo Park ● Data sharing commitment
● Cross jurisdictional collaboration potential