big data analytics for active transportation and …...teresa tapia [email protected]...
TRANSCRIPT
Teresa Tapia
Customer Success Manager
StreetLight Data
Big Data Analytics for Active Transportation
and Multimodal Planning
-- Proprietary and Confidential -- 2
Overview
I. Who We Are
II. Big Data for Active Transportation
I. Our M2 Initiative
II. How We Capture Active Modes
III. Pilot Project Results
-- Proprietary and Confidential -- 3
Zone Activity
Volume
Origin
Destination
Matrices
StreetLight InSight®
We Provide the Best Big Data Resources and
Processing Software for Transportation
AADT/AAHT
Trip Length
Trip Circuity
Congestion
(Free Flow
Factor)
Aggregate
Home/Work
Locations
Select Link
Trip Duration
Demographics
Trip Speed
Big Data +
Contextual Data
-- Proprietary and Confidential -- 4
What Big Data Are We Working With?Mobile device data from ~23% of US and Canadian adults and ~12% of commercial truck trips
Video shows a subset from Oct 8th, 2017 in San Bernardino, California
-- Proprietary and Confidential -- 6
Our On-Demand Platform Delivers Real-World Transportation
Analytics for Data-Driven Policy and Infrastructure Planning
StreetLight InSight®: The Only On-Demand Platform For
Running Actionable Transportation Analytics
1300+Analyses
Supported
Each Month
-- Proprietary and Confidential -- 7
Why We Launched Our Multimodal Measurement
Initiative: All Modes Count – But Not All Are Counted
-- Proprietary and Confidential -- 8
We Are Currently Engaged in Pilot Projects for M2
Working Group Partners
Spring 2018 Summer 2018 Fall 2018 Winter 2018
V1 Active Mode + Gig Driving
Algorithms Finalized
Bike/Ped Pilot Selection and Deliveries
(Selecting working group members with
calibration data and curiosity!)
Launch of full Active Mode
Metrics in StreetLight InSight
Algorithm and product tweaks, based on partner feedback
Gig Driving Pilot Deliveries for Working Group
Big Data Resources for
Active Transportation
-- Proprietary and Confidential -- 10
What’s Out There?
Cell Tower Vehicle / Truck –
Navigation-GPS
Multi-App
Location-Based
Services Data
Mode-
Specific App
Data
Inroad Sensors Video Readers
What
Modes?
Air, some car Car, truck Air, car, truck, bike,
pedestrian, air,
TNC, boat, train,
bus, etc.
Mode that app
is “for”
Bike, car Car, bike, ped
What
Patterns
?
OD, ~count OD, route, speed OD, route, speed,
demographics,
~count, tours
OD, route,
~tours
Count, RT
Presence
Count, RT
Presence
-- Proprietary and Confidential -- 11
Multimodal Planning Still Needs Car Data!
-- Proprietary and Confidential -- 12
But Let’s Talk About the Harder Stuff
-- Proprietary and Confidential -- 13
Active Mode Data Options: Which to Choose?
Multi-App Locational Data Mode-Specific App
Data
Inroad Sensors Video Readers
Pros? “Always all” (all trip types
and modes), large
reprsentative sample size,
OD/route, demographics,
trip purpose, frequency
Mode certainty,
OD/route, good
sample size
Complete count (in
theory), operational
applications
Complete count (in
theory), operational
applications, view many
types of events
Cons? Mode probabalistally inferred User must turn on,
skewed user group
Expense + maintenance,
extendibility, limited
metrics, bikes/peds roam
Expense, extendibility,
limited metrics, occlusion
-- Proprietary and Confidential -- 14
Active Mode Data Options: Which to Choose?
Multi-App Locational Data Mode-Specific App
Data
Inroad Sensors Video Readers
Pros? “Always all” (all trip types
and modes), large
reprsentative sample size,
OD/route, demographics,
trip purpose, frequency
Mode certainty,
OD/route, good
sample size
Complete count (in
theory), operational
applications
Complete count (in
theory), operational
applications, view many
types of events
Cons? Mode probabalistally
inferred
User must turn on,
skewed user group
Expense + maintenance,
extendibility, limited
metrics
Expense, extendibility,
limited metrics, occlusion
We Choose: All of the Above
-- Proprietary and Confidential -- 15
We’re Developing Partnerships with Mode-
Specific Smartphone Apps
-- Proprietary and Confidential -- 16
Machine Learning: Speed is Not Adequate!
-- Proprietary and Confidential -- 17
How We’re Putting Everything Together
+B
A• 40 on Road Z • 60 on Road Z
1000 Bike Trip Counts
• 100 on Road Z
TOTAL
Pilot Project Results
-- Proprietary and Confidential -- 19
Completed Pilot Project – Active O-D to Transit
Stations in Sacramento
Implication for policy
makers and planners:
A large portion of people
walk and bike to and from
areas across the freeway.
Prioritize connectivity
across the freeway for
pedestrians / bike bridges.
bit.ly/ConnectingSacramentoOrigins and destinations of active trips to and from Zinfandel Station and
Cordova Town Center
Zinfandel Station
Cordova Town Center Station
-- Proprietary and Confidential -- 20
Completed Pilot Project – Campus Circulation
Study for University of Miami in Ohio
Two main arterials, US 27
and SR 73, bring regional
traffic through campus
Implication for University’s
Transportation Planning:
Shows where additional
cross walks and improved
signal timing are most
needed to improve flow of
vehicle traffic during class
exchange times.
Active transportation (bike + ped
combined) data showed 46% of
active trips cross US 73 into Zone
1