Taking Pedestrian and Bicycle Counting Programs to the Next Level

MOVINGFORWARDTHINKING
Taking Pedestrian and Bicycle Counting
Programs to the Next Level
Lessons from NCHRP 07-19
Robert Schneider, UW Milwaukee
RJ Eldridge, Toole Design Group
Conor Semler, Kittelson & Associates
ProWalk/ProBike/ProPlace
Pittsburgh, PA
September 2014 Source: Kittelson & Associates, Inc.

Presentation Overview
 Introduction
 State of the Practice
 Count Applications
 Testing Approach and
Findings
 Guidebook Overview
 Questions and
Discussion
2
Source: Toole Design Group

NCHRP 7-19 Research Team
 Kittelson & Associates, Inc.
– Principal Investigator: Paul Ryus
 University of Wisconsin-Milwaukee
 UC Berkeley, SafeTREC
 Toole Design Group
 McGill University
 Quality Counts, LLC
3

Project Purpose
 Address lack of
pedestrian and
bicycle volume data
 Assess data
collection
technologies and
methods
 Develop guidance
for practitioners
4
Source: Bob Schneider, UW-Milwaukee

Related Work
 National Bicycle and Pedestrian
Documentation Project
 FHWA Traffic Monitoring Guide
(TMG)
– 2013 edition includes chapter on
non-motorized traffic
– NCHRP research complements
FHWA guide
5

State of the Practice
 Early Findings
 Multimodal Count Applications
6
Source: Bob Schneider, UW-Milwaukee

NCHRP 7-19 Survey (N = 269)
10
20
36
3
11
69
2
35
39
20
0 10 20 30 40 50 60 70 80
Other
University
STATE DOT
U.S. federal agency
U.S. County
U.S. city
Transit agency
Non-profit/advocacy
MPO
Consulting firm
SURVEY RESPONSES

NCHRP 7-19 Survey Findings
• Pedestrian and bicycle counts are becoming routine
for cities, MPOs, and State DOTs.
• Manual counts are the most prevalent data collection
method
• Most programs lack formal or dedicated funding
source and rely heavily on volunteers
• Organizations serving larger communities are more
likely to conduct pedestrian and/or bicycle counts.
• Some integration with motor vehicle count databases.

Current Methods of Bicycle Counting

NCHRP 7-19 Survey Findings
 There is no standard approach for count
programs
 Practitioners are looking for more guidance
– Choosing devices
– Selecting locations
– Count intervals and duration
– Temporal/seasonal adjustments

Pedestrian & Bicycle Count Applications
 Measuring facility usage
 Evaluating before-and-after data
 Monitoring travel patterns
 Safety analysis
 Project prioritization
 Multimodal modeling
11

Measuring Facility Usage
 Transportation system monitoring program
 Typically requires collecting counts at set
locations and regular intervals
 Critical for tracking progress and measuring
success
12
Change in Walking and
Bicycling Activity at
Washington State Count Sites,
2009–2012
Source: Washington State DOT (2012)

Evaluating Before-and-After Volumes
 Measure volumes before and after facility is
opened
 Forecast usage of planned facilities
13
Before-and-After Bicycle Facility Usage – buffered bicycle lanes on Pennsylvania Ave
Source: Kittelson &
Associates, Portland State
University, and Toole Design
Group (2012)

Monitoring Travel Patterns
 Developing extrapolation factors
– Extrapolate short-duration counts over longer time
periods
– Control for effect of land use, weather,
demographics, etc.
 Evaluating user behavior patterns
– Identify factors that influence walking/biking
– Controllable (land use) and uncontrollable (weather)
factors
14

Safety Analysis
 Quantifying exposure
– Variety of methods proposed to quantify exposure
– One method compares pedestrian-vehicle collissions
to average annual pedestrian volumes
 Identifying before-and-after safety effects
15

Mainline
Roadway
Intersecting
Roadway
Reported
Pedestrian
Crashes (1996-
2005)
Mission
Boulevard
Torrano
Avenue 5
Davis Street Pierce Avenue 4
Foothill
Boulevard D Street 1
Mission
Boulevard
Jefferson
Street 5
University
Avenue Bonar Street 7
International
Boulevard 107th Avenue 2
San Pablo
Avenue Harrison Street 2
East 14th
Street
Hasperian
Boulevard 1
International
Boulevard 46th Avenue 3
Solano Avenue
Masonic
Avenue 2
Broadway 12th Street 5
Alameda County Pedestrian Crash Analysis

Mainline Roadway
Intersecting Roadway
Estimated Total Weekly Pedestrian Crossings
Annual Pedestrian Volume Estimate
Ten-Year Pedestrian Volume Estimate
Reported Pedestrian Crashes (1996- 2005)
Pedestrian Risk (Crashes per 10,000,000 crossings)
Mission Boulevard
Torrano Avenue
1,169
60,796
607,964
5
82.24
Davis Street
Pierce Avenue
1,570
81,619
816,187
4
49.01
Foothill Boulevard
D Street
632
32,862
328,624
1
30.43
Mission Boulevard
Jefferson Street
5,236
272,246
2,722,464
5
18.37
University Avenue
Bonar Street
11,175
581,113
5,811,127
7
12.05
International Boulevard
107th Avenue
3,985
207,243
2,072,429
2
9.65
San Pablo Avenue
Harrison Street
4,930
256,357
2,563,572
2
7.80
East 14th Street
Hasperian Boulevard
3,777
196,410
1,964,102
1
5.09
International Boulevard
46th Avenue
12,303
639,752
6,397,522
3
4.69
Solano Avenue
Masonic Avenue
22,203
1,154,559
11,545,589
2
1.73
Broadway
12th Street
112,896
5,870,590
58,705,898
5
0.85

Mainline Roadway
Intersecting Roadway
Estimated Total Weekly Pedestrian Crossings
Annual Pedestrian Volume Estimate
Ten-Year Pedestrian Volume Estimate
Reported Pedestrian Crashes (1996- 2005)
Pedestrian Risk (Crashes per 10,000,000 crossings)
Mission Boulevard
Torrano Avenue
1,169
60,796
607,964
5
82.24
Broadway
12th Street
112,896
5,870,590
58,705,898
5
0.85

Project Prioritization
 Identify high-priority locations
for improvements
 Identify factors that influence
walking/biking and prioritize
accordingly
 Measure improper user
behaviors (i.e., wrong-way bike
riding) to identify areas
needing improvement
19

Multimodal Model Development
 Multimodal travel
demand modeling is an
emerging field
 Potential to estimate
demand over a large
area and forecast
influence of
infrastructure changes
20
Source: City of Berkeley, CA
Pedestrian Master Plan

We need these data fields!
Institutionalize Pedestrian & Bicycle Data
 A multimodal transportation system requires
collecting data for all modes of transportation
 Establish baseline for pedestrian & bicycle safety,
infrastructure, volumes, etc.
X-Street
Traffic
Volume
Mainline
Traffic
Volume
X-Street
Pedestrian
Volume
Mainline
Pedestrian
Volume

Challenges to Bicycle and Pedestrian
Counting
 Where to count?
 When (how long/how
frequently) to count?
 What to count?
 How to count?
 Site/Mode challenges
 Who should be counting?

Arlington County’s automated
bicycle and pedestrian count
program
• First counter Custis Trail:
Fall 2009
• Now 30 automatic counter
stations
• Automatic uploads to
central server
• Web based “dashboard”
• Plans for real-time “totem”
counter display
Case Study: Arlington County, VA
Statistics and graphics courtesy of David Patton
Bicycle & Pedestrian Planner, Arlington County Division
of Transportation

NCHRP 7-19 Research Approach
 Focus on testing and evaluating commercially
available automated technologies
 Assess type of technology as opposed to a
specific product
 Cover a range of facility types, mix of traffic,
and geographic locations
 Evaluate accuracy through the use of manual
count video data reduction
24

Motor Vehicle Data Collection
 Inductive Loops
 Pneumatic Tubes
 Manual Count Boards
 Video cameras
 ITS integration
 In-vehicle sensors (toll
tags)
Photo Credit: Federal Highway Administration

Auto versus Multimodal Counts
 Key differences:
– Differences in demand
variability
– Ease of detection
– Experience with
counting technology
27
0
2,000
4,000
6,000
8,000
10,000
12,000
0 6 12 18 24
Hourly Volume (veh/h)
Hour Starting
Monday Tuesday Wednesday Thursday Friday Annual Weekday Average

Motor Vehicle Data Collection
Constrained; somewhat
predictable

Bicycle Data Collection
Constrained environments
easy to monitor
Complex environments
harder to define
Detection

Pedestrian Data Collection
Constrained environments
easy to monitor
Detection
People tend to make
their own path

Motor vehicle data
collection
• Widely collected
• Easy to track vehicle
movements
• Predictable patterns and
routes
• Years of trend data to
analyze
Bicycle and pedestrian
data collection
• Sparsely collected
• Difficult to track and
tabulate movements
• Unpredictable paths of
travel
• Weather and seasonal
impacts
• Lack of historical data
Challenge: Site/Mode characteristics

Counting System
 Sensor technology one piece of
counting system
32

NCHRP 07-19 Evaluation
Automated devices that count all users
(pedestrians and bicyclists)

Passive Infrared (IR)
 Detect pedestrians and
cyclists by infrared
radiation (heat) patterns
them emit
 Passive infrared sensor
placed on one side of
facility
 Widely used and tested
35

Active Infrared (IR)
 Transmitter and
receiver with IR beam
 Counts caused by
“breaking the beam”
36
Source: Steve Hankey, University of Minnesota

Radio Beam
 Radio signal between
transmitter and receiver
 Detections occur when
beam is broken
 Not previously tested in
literature
 Some distinguish bikes
from peds
37
Source: Karla Kingsley, Kittelson & Associates, Inc.

Automated devices that count
bicyclists only

Pneumatic Tubes
 Tube(s) stretched across
roadway
 When a bicycle rides over
tube, pulse of air passes
through tube to detector
39
Source: Karla Kingsley, Kittelson & Associates, Inc.

Inductive Loops
 Generate a magnetic field
that detect metal parts of
bicycle passing over loop
 In-pavement or
temporary surface loops
40

Piezoelectric Sensor
 Emit an electric signal
when physically deformed
 Typically embedded in
pavement across travel
way
41

Other counting methods

Combination
 Use one technology to detect all others plus
another technology to detect bicyclists only
 Provide bicycle and pedestrian counts
separately
43
Photo Sources: Bob Schneider, UW-Milwaukee

Manual Counts
 Most common type of counting, to date
 Capture many different locations
 Record pedestrians & bicyclists separately
 Can count road crossings
 Capture user characteristics
(gender, helmet use,
behavior, etc.)
Photo by Robert Schneider, UW-Milwaukee

Washington State DOT Pedestrian & Bicycle Counts
Source: Cascade Bicycle Club. Washington State Bicycle and Pedestrian Documentation Project, Prepared for the Washington State Department
of Transportation, February 2013.

Existing Market Technology
• Passive infrared
• Active infrared
• Radio beam
• Pneumatic tubes
• Inductive loops
• Piezoelectric sensors
• Combination devices
• Manual Counts
• Automated video
• Emerging technologies
Understanding Count Technologies
Classify Bicycle Only Everything
Sources: Toole Design Group

Untested/Emerging technologies
 Thermal
 Radar
 Laser scanners
 Pressure and acoustic sensors
 Automated video
47

Related Work
 Topics related to quantifying pedestrian & bicycle
activity, but not covered:
– Bluetooth and WiFi detection
– GPS data collection
– Cell/smartphone data collection
– Radio frequency ID (RFID) tags
– Bike sharing data
– Pedestrian signal actuation buttons
– Surveys
– Presence detection
– Trip generation
48

Quick Break for Questions
49
Detection

Study Technologies and Site Locations
 Technologies
– Passive infrared
– Active infrared
– Pneumatic tubes
– Inductive loops
– Piezoelectric
– Radio beam
– Combination of
technologies
50
 Site Locations
– Portland, OR
– San Francisco, CA
– Davis, CA
– Berkeley, CA
– Minneapolis, MN
– Washington, D.C.
– Arlington, VA
– Montreal, Canada

Washington, D.C. and Arlington, VA
 Key Bridge separated path
– Passive Infrared
– Pneumatic Tubes
– Passive Infrared with inductive loops
51
Source: Toole Design
Group

Minneapolis, MN
 Midtown Greenway
multiuse path
– Active Infrared
– Radio Beam
– Inductive Loops
– Pneumatic Tubes
52

 Eastbank Esplanade
multiuse path
– Pneumatic Tubes
– Radio Beam
Portland, OR
53
Source: Kittelson & Associates, Inc.

San Francisco, CA
 Fell Street Bicycle Lane
– Pneumatic Tubes
– Inductive Loops
54
Source: Frank Proulx, UC Berkeley SafeTREC

Evaluation Method
 Video-based manual counts
 Interrater reliability tested
55

Measures to Evaluate the Quality of Count
Data from Technologies
 Accuracy: The magnitude of the difference
between the count produced by the technology,
and actual (“ground-truth”) count.
(Typically depends on user volumes, movement patterns, traffic mix, and environmental
characteristics)
 Consistency (Precision): The remaining variability in
the count data after being corrected for expected
under- or over-counting, given specific conditions.
(Typically depends on the counting technology itself, how a specific vendor uses the
technology in a particular product, and quality of installation)
56

Graphical Analysis
57
Undercounting
Overcounting

Active Infrared Counter Validation Data
(Somewhat accurate, very precise)

(Somewhat accurate, but not precise)
Passive Infrared Counter Validation Data

Evaluation Criteria
 Accuracy
 Precision
 Ease of installation, maintenance
requirements, cost, flexibility of data, etc.
60

Summary of Data Collected
Condition
Passive
Infrared
Active
Infrared
Pneumatic
Tubes
Inductive
Loops
Inductive
Loops
(Facility)
Piezo-electric
Radio
Beam
Total hours of data 298 30 160 108 165 58 95
Temperature (°F)
(mean/SD)
70 / 15 64 / 26 71 / 9 73 / 12 71 / 17 72 / 10 74 / 10
Hourly user volume
(mean/SD)
240 / 190 328 / 249 218 / 203 128 / 88 200 / 176 128 / 52 129 / 130
Nighttime hours 30 3 10 13 19 15.75 3.5
Rain hours 17 0 4 7 7 0 6
Cold hours (<30 °F) 12 5 0 0 7 0 0
Hot hours (>90 °F) 11 0 0 5 5 3 4
Thunder hours 8 0 0 2 2 0 0
61

Passive Infrared
 Easy installation
 Mounts to existing pole/surface or in purpose-built
pole
 Potential false detections from background
 Possible undercounting due to occlusion
62

Passive Infrared
 Average undercount rate
8.75%
 Differences between
products tested
 Correction function could
account for facility width
 Accuracy not affected by
high temperatures
63

Active Infrared
 Moderately easy
installation – requires
aligning transmitter and
receiver
 Single device tested –
accurate and highly
precise
64

Pneumatic Tubes
 Tested BSCs – bicycle specific counters
 Primarily tested tubes on multi-use paths and
bicycle lanes
 Issues with site on 15th Avenue in Minneapolis
– Tube nails came out of ground
– Severe undercounting and overcounting
– Removed sensors from data set
65

Pneumatic Tubes
 Fairly high accuracy at
very high volumes
 Accuracy rates not
observed to decline with
aging tubes
 Future research in mixed
traffic settings
66

Radio Beam
 Required mounting devices 10’ apart
 Tested two products: one that distinguished
bicyclists and pedestrians (product A), one that
did not (product B)
67

Radio Beam
 Product B higher
accuracy
 Product A – low
precision and
lower accuracy
 Occlusion errors
 Temperature,
lighting, rain
issues
68

Inductive Loops
 Permanent (in ground) or temporary (on
surface)
 Bypass errors
– Cyclists passing
outside bike lane
– Loops leaving gaps
in detection zone
69

Inductive Loops
 Accurate and precise
 Errors with age of
loops not detected
 Higher volumes
slightly affect
accuracy
 No substantial
difference between
permanent and
temporary loops
70

Inductive Loops
 Need to mitigate bypass errors
71

Piezoelectric Strips
 Tested one existing
device, due to
difficulties procuring
equipment
 Data suggests device is
not functioning very
precisely or accurately
 Caution – data from
single device not
installed by research
team
72

Combination Counter
 Passive infrared +
inductive loops
 Each device also
assessed separately
 Inferred pedestrian
volumes (Total – bikes)
 Occlusion errors and
undercounting
 High rate of precision
73

Accuracy Calculations
74
 Average percentage deviation (APD): overall
divergence from perfect accuracy
(undercounting rate)
 Average of the absolute percentage deviation
(AAPD): accounts for undercount/overcount
cancelation (total deviation)

Research Conclusions
Device Undercounting Rate Total Deviation
Passive Infrared (2 products) 8.75% 20.11%
Active Infrared 9.11% 11.61%
Pneumatic Tubes 17.89% 18.50%
Radio Beam 18.18% 48.15%
Inductive Loops 0.55% 8.87%
Piezoelectric Strips 11.36% 26.60%
75
 Automated counter accuracy:

Research Conclusions
 Factors influencing accuracy
– Proper calibration and installation
– Occlusion
– Vendor differences
 Factors not found to influence accuracy
– Age of inductive loops or pneumatic tubes
– Temperature
– Snow/rain (limited data)
76

Guidebook Organization
Quick Start Guide
1. Introduction
2. Non-Motorized Count Data Applications
3. Data Collection Planning and Implementation
4. Adjusting Count Data
5. Sensor Technology Toolbox
Case Studies
Manual Pedestrian and Bicyclist Counts: Example Data Collector
Instructions
Count Protocol Used for NCHRP Project 07-19
Appendix D. Day-of-Year Factoring Approach
77
Appendices

2. Non-Motorized Count Applications
 Measuring facility usage
 Evaluating before-and-after data
 Monitoring travel patterns
 Safety analysis
 Project prioritization
 Multimodal modeling
Source: Kittelson & Associates,
Portland State University, and Toole
Design Group (2012)
Before-and-After Bicycle Facility Usage –
buffered bicycle lanes on Pennsylvania Avenue
For each
application:
Details
Case Studies
78

3. Data Collection Planning & Implementation
 Covers:
1. Planning the count
program
2. Implementing the
count program
 Provides examples,
detailed guidance,
checklists
Source: Toole Design Group.
79

Common Strategy: Manual + Automated
Geographic Coverage
A Few
Locations
Many
Locations
Count Duration
Continuous Automated
Short-Term Manual

Comparison of Counting Technologies
Characteristic
Passive
Infrared
Active
Infrared
Pneumatic
Tubes
Inductive
Loops
Piezoelectric
Sensor
Passive IR +
Inductive
Loops
Radio Beam
(One
Frequency)
Radio Beam
(High/Low
Frequency)
Automated
Video1
Manual
Counts2
Type of users counted
All facility users Yes Yes Yes Yes Yes Yes Yes
Pedestrians only Yes Yes Yes Yes
Bicycles only Yes Yes Yes Yes Yes Yes Yes
Pedestrians vs. bicycles Yes Yes Yes Yes
Bicycles vs. automobiles Yes Yes Yes Yes
Characteristics collected
Different user types Yes Yes Yes Yes
Direction of travel3 Yes Yes Yes Yes Yes Yes Yes Yes Yes
User characteristics4 Yes Yes
Types of sites counted
Multiple-use trail segments Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Sidewalk segments Yes Yes Yes Yes Yes Yes Yes
Bicycle lane segments Yes Yes Yes Yes Yes
Cycle track segments Yes Yes Yes Yes Yes Yes
Shared roadway segments Yes Yes Yes Yes
Roadway crossings (detect
from median)5 Yes Yes Yes Yes Yes Yes Yes Yes
from end of crosswalk) Yes Yes
Intersections (identify
turning movements) Yes
Notes: (1) Existing “automated video” systems may not use a completely automated counting process; they may also incorporate manual data checks of automated video processing.
(2) Includes manual counts from video images.
(3) Technologies noted as “Yes” have at least one vendor that uses the technology to capture directionality.
(4) User characteristics include estimated age, gender, helmet use, use of wheelchair or other assistive device, pedestrian and bicyclist behaviors, and other characteristics.
(5) Roadway crossings at medians potentially have issues with overcounting due to people waiting in the median. Median locations were not tested during this project.

Characteristic
Passive
Infrared
Active
Infrared
Pneumatic
Tubes
Inductive
Loops
Piezoelectric
Sensor
Passive IR +
Inductive
Loops
Radio Beam
(One
Frequency)
Radio Beam
(High/Low
Frequency)
Automated
Video1
Manual
Counts2
Type of users counted
All facility users Yes Yes Yes Yes Yes Yes Yes
Pedestrians only Yes Yes Yes Yes
Bicycles only Yes Yes Yes Yes Yes Yes Yes
Pedestrians vs. bicycles Yes Yes Yes Yes
Bicycles vs. automobiles Yes Yes Yes Yes
Characteristics collected
Different user types Yes Yes Yes Yes
Direction of travel3 Yes Yes Yes Yes Yes Yes Yes Yes Yes
User characteristics4 Yes Yes
Types of sites counted
Multiple-use trail segments Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Sidewalk segments Yes Yes Yes Yes Yes Yes Yes
Bicycle lane segments Yes Yes Yes Yes Yes
Cycle track segments Yes Yes Yes Yes Yes Yes
Shared roadway segments Yes Yes Yes Yes
from median)5 Yes Yes Yes Yes Yes Yes Yes Yes
from end of crosswalk) Yes Yes
Intersections (identify
turning movements) Yes
Notes: (1) Existing “automated video” systems may not use a completely automated counting process; they may also incorporate manual data checks of automated video processing.
(2) Includes manual counts from video images.
(3) Technologies noted as “Yes” have at least one vendor that uses the technology to capture directionality.
(4) User characteristics include estimated age, gender, helmet use, use of wheelchair or other assistive device, pedestrian and bicyclist behaviors, and other characteristics.
(5) Roadway crossings at medians potentially have issues with overcounting due to people waiting in the median. Median locations were not tested during this project.
Comparison of Counting Technologies

INSTALLATION CHECKLIST: ADVANCE PREPARATION
Site visit to identify the specific installation location. Specifically, note poles that will be
used, where pavement will be cut, or where utility boxes will be installed to house
electronics. Verify that no potential obstructions (e.g., vegetation) or sources of
interference (e.g., doorway, bus stop, bicycle rack) are present.
Obtain and document necessary permissions. Permits or permissions may include right-of-
way encroachment permits, pavement cutting permits or bonds, landscaping
permits, or interagency agreements. Obtaining these permissions may take up to
several months, particularly if other agencies are involved.
Create a site plan. Develop a detailed diagram of the planned installation on an aerial
photo or ground-level image documenting the intended equipment installation
locations and anticipated detection zone (after installation this will be useful for
validating equipment either visually or with video monitoring). This diagram may be
useful for obtaining installation permissions and working with contractors. Figure 3-6
provides an example site plan.
Hire a contractor if necessary (or schedule appropriate resources from within the
organization).
Arrange an on-site coordination meeting involving all necessary parties (e.g., staff
representing the organization installing the counter, permitting staff, contractors). If
possible, a vendor representative should be on hand or available by phone. It may
take several weeks to find a suitable time when everyone is available.
Check for potential problems. Problems with the site may include interference from
utility wires, upcoming constructions projects, hills, sharp curves, nearby illicit
activity, and nearby insect and animal activity. Some of these conditions can be
identified from imagery, but they should also be evaluated in the field.

EQUIPMENT MONITORING CHECKLIST
Sensor height. Is the sensor still mounted at the correct height?
Sensor direction. Is the sensor still pointed in the correct direction?
Clock. Is the device’s clock set correctly?
Cleaning. Remove dirt, mud, water, or other material that may affect the sensor or other
vital components.
Battery. Is the remaining battery life sufficient to last until the next scheduled visit?
Obstructions. For example, is vegetation growing too close to the device?
Unanticipated site problems. For example, is the pole being used for bicycle parking, or
are people congregating in the area (as opposed to walking past the counter)?
Pedestrian or bicycle detection. Are pedestrians or bicyclists passing through the
counter’s detection zone being counted? If not, can the counter’s sensitivity be
adjusted in the field, or does it need to be removed for repairs?
Download data. Use the same export option consistently to ease the data management
burden back in the office.
Securement. Are the installation elements and locking devices still secure and durable?
Poorly secured or loose fitted devices are more vulnerable to theft and vandalism.

4. Adjusting Count Data
 Sources of counter inaccuracy
 Measured counter accuracy
 Counter correction factors
 Expansion factors
 Example applications
Occlusion error
92

Linear
Correction Factors
Table 4-2. Simple Counter Correction Factors Developed by NCHRP Project 07-19
Sensor Technology
Adjustment
Factor Hours of Data
Passive infrared 1.137 298
Product A 1.037 176
Product B 1.412 122
Active infrared† 1.139 30
Radio beam 1.130 95
Product A (bicycles) 1.470 28
Product A (pedestrians) 1.323 27
Product B 1.123 39
Pneumatic tubes* 1.135 160
Product A 1.127 132
Product B 1.520 28
Surface inductive loops* 1.041 29
Embedded inductive loops* 1.054 79
Piezoelectric strips† 1.059 58
Notes: *Bicycle-specific products.
†Factor is based on a single sensor at one site; use caution when applying.

5. Treatment Toolbox
 Description
 Typical application
 Level of effort
 Strengths
 Limitations
 Accuracy
 Usage
Sidebar with
quick facts
96

Summary: Key Issues for Practice
 Consider automated + manual counts
 Determine whether to use permanent or mobile
counters (or both)
 Accuracy of devices & service vary by vendor
 Results are useful for general corrections, but best
practice is to conduct local ground-truth counts &
make specific corrections
 Consider approvals and site characteristics when
selecting a count site
 How critical is accuracy?
97

Suggested Research
 Additional testing of automated technologies
– Technologies not tested or underrepresented
– Additional sites and conditions
 Extrapolating short-duration counts to longer-duration
counts
 Adjustment factors for environmental factors
 Uniform approach(es) for FHWA TMG?
98

Questions and Discussion
 Contact Information
– Conor Semler, Kittelson & Associates, Boston, MA
csemler@kittelson.com, 857.265.2153
– Robert Schneider, UW-Milwaukee, Milwaukee, WI
rjschnei@uwm.edu, 414.229.3849
– RJ Eldridge, Toole Design Group, Washington, DC
reldridge@tooledesign.com, 301.927.1900 x107
99

Plan the Count Program
 Specify the data collection purpose
 Identify data collection resources
 Select count locations & determine timeframe
 Consider available counting methods

Implement the Count Program
 Obtain necessary permissions
 Procure counting devices
 Inventory and prepare devices
 Train staff
 Install and validate devices
 Calibrate devices
 Maintain devices
 Manage count data
 Clean and correct count data

Three Types of Data Adjustments
 Cleaning erroneous data involves identifying when a count is likely to
represent a time period when an automated counter was not
observing the intended pedestrian or bicycle movements. Data from
these time periods are discarded or replaced by better estimates.
 Correcting for systematic errors involves applying a correction
function that removes the expected amount of under- or
overcounting for a particular counting technology (often due to
occlusion).
 Expanding short counts to longer time periods involves applying an
expansion factor to a count collected for a short time period. The
expansion factor is based on the type of activity pattern assumed to
exist at the site, and it can account for the effects of weather on
pedestrian and bicycle volumes.

Weekday
Hour Raw Count
Cleaned
Count
Corrected
Count
Extrapolated
Count
0:00 10 10 11.2 11.2
1:00 5 5 5.6 5.6
2:00 3 3 3.36 3.36
3:00 2 2 2.24 2.24
4:00 3 3 3.36 3.36
5:00 15 15 16.8 16.8
6:00 65 65 72.8 72.8
7:00 125 125 140 140
8:00 150 150 168 168
9:00 140 140 156.8 156.8
10:00 95 95 106.4 106.4
11:00 105 105 117.6 117.6
12:00 130 130 145.6 145.6
13:00 0 115 128.8 128.8
14:00 100 100 112 112
15:00 125 125 140 140
16:00 140 140 156.8 156.8
17:00 160 160 179.2 179.2
18:00 155 155 173.6 173.6
19:00 145 145 162.4 162.4
20:00 120 120 134.4 134.4
21:00 85 85 95.2 95.2
22:00 190 45 50.4 50.4
23:00 20 20 22.4 22.4
0:00 11.2
1:00 5.6
2:00 3.36
3:00 2.24
4:00 3.36
Three Types of Data
Adjustments
1 2 3

Example: 24-hours on one weekday

Adjustment 2: Correct Data
+X% to correct for
undercounting

Raw Count Data with Missing Data Imputed

Taking Pedestrian and Bicycle Counting Programs to the Next Level

More Related Content

What's hot

Similar to Taking Pedestrian and Bicycle Counting Programs to the Next Level

More from Project for Public Spaces & National Center for Biking and Walking

Recently uploaded

Taking Pedestrian and Bicycle Counting Programs to the Next Level