Data Driven Design - Matthew Roe, New York City DOT

Data-Driven Design: State-of-the-Art Street Engineering

Istanbul 26 August 2013

•  Urban Context & Challenge

•  Pedestrian Safety Metrics

•  Projects/Design

•  Conclusions future research 2

Today’s Presentation

6,300 mi of streets & highways 781 bridges, 6 tunnels 12,000+ signalized intersections

•  Needs –  Measure city’s performance –  Prioritize projects/measure facility performance –  Evaluate projects/design changes

3

Measurement Challenges

•  Challenges –  “NYC is different” –  Fairness to modes – severity, exposure –  Precision, timeliness –  Work In Real Life

•  Speed & Arterials •  Likelihood of KSI increases steeply with

increased speed •  Late-night crashes twice as deadly for

pedestrians •  2/3 of pedestrian fatalities are on arterial streets

(<15% of network) •  Deadly combination: Speeding + midblock/

against signal crossing

•  Left Turns & Other Conflicts –  Both one-way and two-way arterials –  3 times as many pedestrian KSI as right turns –  47% of pedestrians killed in crosswalks had

right-of-way.

4

Pedestrian Safety Challenges

Academic study of 5 years of pedestrian KSI cases (NYU/RPI/SUNY-Buffalo)

•  National fatality rates are >3x higher than NYC rates •  NYC has low PMT & VMT per capita

–  Safer per trip, but more risky per mile. •  Single-mode stats and PMT/VMT stats punish dense, safe urban areas

5

Exposure Matters for City Comparisons

Traffic Fatalities per 100,000 Residents Yearly Average (2008-2010)

Journey-to-Work Transit + Walking

Mode Share (2008-2010)

Pedestrian Non-Pedestrian Total

NYC 1.8 1.4 3.3 68.3%

Peer Cities 2.1 4.4 6.5 26.8%

USA (less NYC) 1.4 10.2 11.5 8.2%

Sources: NYCDOT, NHTSA FARS, Census ACS 2010 3-year estimates (excl. worked at home)

6

•  Divide bicyclist KSI by indexed commuter bike volume

•  Show improvement in safety per cyclist

•  Not a comparison to other modes

•  Risk indicator important when use of a particular mode is changing rapidly

•  Similar method could be used for specific facilities/sub-city areas if data is available

Exposure Matters for Growing Modes

•  KSI per mile over 5-year period (all modes) –  KSI: Persons Killed or Severely Injured.

(Fatalities + “A” Injuries)

–  Vulnerable road users more represented than in total injuries

–  More spatiotemporal consistency than fatalities, but similar crash characteristics

–  Reflects crash severity without requiring a specific weighting system

•  High Crash Corridors: top 1/3 of mileage in each borough

•  Allows quick project prioritization –  Corridor safety issues should be addressed at

the corridor level –  Fair to all modes –  KSI/mile represents problem/cost

7

Prioritizing Corridors

8

Information for Planners and Designers

•  KSI by mode

•  High-Crash Corridor designation

•  Injuries by severity by mode

•  Map of corridor •  Interface allows project managers to access

safety data quickly, determine priority level

•  High Crash Locations prioritized

•  Involve communities & local leaders at depth commensurate with project

•  In-house implementation

•  Fast – 6 to 18 months from initiation to completion

9

Planning & Implementation

Monitoring & Before-After •  Injury crashes •  KSI •  Radar speeds •  Other data:

–  Travel Times and/or LOS

–  Community feedback

–  Economic development

10

Evaluation

•  Before: –  60’ –  Two lanes each

direction

•  After: –  One lane each

direction –  Left turn bays –  Bike lanes or wide

parking lanes (13’) –  Planted refuge

islands

11

Project Type: 4-to-3-Lane Conversion

Empire Boulevard, Brooklyn

•  Before: – 50’, two lanes

•  After: – Flush center

median – Left turn bays – Bike lanes or

wide parking lanes

12

Project Type: 2-to-3-Lane Conversion

E. 180th Street, Bronx

13

Project Type: Protected Bike Path

•  Before: –  Multi-lane one-way –  Marked bicycle lane

•  After: –  Parking-protected

on-street bike path –  Parking lane or

pedestrian plaza –  Left turn bays w/

signal or “mixing zones”

–  Shortened crossings 9th Avenue, Manhattan

•  Before: –  Mixed traffic –  Lefts vs.

pedestrians

•  After: –  Dedicated left

turn signal from avenue

–  Shortened crossings

14

Project Type: Left Turn Separation

7th Avenue & W. 23rd Street, Manhattan

•  Before: –  100’ wide –  Narrow/short islands –  Narrow LT lane –  3 lanes, narrow right

lane

•  After: –  Wider pedestrian

islands –  2 standard lanes –  Wide parking lane –  Longer/wider LT

lanes

15

Project Type: Center Median Widening

4th Avenue, Sunset Park, Brooklyn

•  Before: – Narrow medians

don’t extend into crosswalk

•  After: – Widened median

tips extend into crosswalk as pedestrian islands

16

Project Type: Arterial Street Median Tips

Queens Boulevard, Queens

•  Speed limit lowered to 20 mph (from 30)

•  Small, self-contained area

•  Announced with signs and gateway treatments

•  Self-enforcing speed humps, parking lane stripes, standardized lane widths

17

Project Type: Neighborhood Slow Zones

Claremont Slow Zone, Bronx

•  Measuring Safety –  Use risk per person as comprehensive performance measure –  Use exposure measures useful for fast-growing modes, locations –  KSI or weighted crashes are fair metrics for urban streets –  Need user-friendly data at corridor or intersection level –  Prioritize loosely (strict rankings don’t work in real life)

•  Designing with Data –  Problems: Speed, left turns, crossing against signal & midblock –  Designs: removing extra lanes, organizing left turns, reducing time/distance

between crossings –  Implement quickly, learn quickly

18

Conclusions

Questions

Matthew Roe matthewjafferoe@gmail.com

v

Decade NYC

Pedestrian Fatalities

Avg per Year

Pedestrian Fatalities per

100,000 Residents per Year

% Pedestrian

1910 – 1919 381 7.3 70%

1920 – 1929 735 11.7 70%

1930 – 1939 693 9.6 70%

1940 – 1949 567 7.4 84%

1950 – 1959 454 5.8 72%

1960 – 1969 434 5.5 60%

1970 – 1979 386 5.2 52%

1980 – 1989 331 4.6 55%

1990 – 1999 261 3.4 51%

2000 – 2009 167 2.0 51%

20

The Century

Year Fatalities All Modes

Severe Injuries

All Modes

Pedestrian Fatalities

Pedestrian Severe Injuries

2001 393 5,417 193 1,452

2002 386 5,820 186 1,417

2003 362 5,434 177 1,418

2004 297 4,823 155 1,311

2005 321 4,585 157 1,285

2006 324 4,834 168 1,353

2007 274 4,501 139 1,313

2008 291 4,380 151 1,308

2009 258 4,101 156 1,161

2010 271 4,040 152 1,155

2011 245 4,323* 139 1,160*

21

•  38% reduction in total traffic fatalities since 2001

•  28% reduction in pedestrian fatalities since 2001

•  >20% reduction in pedestrian and all severe injuries since 2001

•  Goal: 50% reduction in all fatalities from 2007 to 2030

The Decade

* Preliminary

•  Population 8.25 m + •  302 mi2 (783 km2)

•  NYC Department of Transportation: –  6,300 mi of streets &

highways –  781 bridges, 6 tunnels –  12,000+ signalized

intersections – Staten Island Ferry – Not subway/bus

operations (MTA)

22

New York City & NYCDOT