ATS-16: Making Data Count, Krista Nordback

Bike-Ped Portal
The National Online Non-motorized Traffic Count Archive
Krista Nordback, Ph.D., P.E., Kristin Tufte, Ph.D.
Morgan Harvey, Nathan McNeil
March 14, 2016
Oregon Active Transportation Summit

Thank you to our partners!
Oregon
Community
Foundation

Introduction to Bike-Ped Portal
• Why?
• What is Bike-Ped Portal?
• How to use it

How many bike and walk?
• Surveys
• National
• Regional
• Local
• Intercept
• GPS
• Counts
• Permanent
• Short duration
• Manual
• Automated
Source: Community Cycles

Why aggregate bicycle and
pedestrian count data?

Bike/Ped Portal
Upload Download
Storage
Data Checking
(QA/QC)
Data Visualization

Bike-Ped Portal
• Online database bp.its.pdx.edu
• 5 million records loaded for 5 states
• Upload/download data

Uploaded Data
• 5 states
• 12 counties
• 343 road or path segments (including 278 in Oregon)
• 355 detectors (both human and machine)
• 38 million people counted
Bicycle
65%
Other
9%
Pedestrian
26%
MODES

Facility Types
Roadway
21%
Path
45%
Sidewalk
7%
Crosswalk
1%
Cycle track
1%
Bike Lane
24%
General
Activity Count
1%
FACILITY TYPES

ATS-16: Making Data Count, Krista Nordback

0
100
200
300
400
500
600
700
800
12:00:00AM
2:00:00AM
4:00:00AM
6:00:00AM
8:00:00AM
10:00:00AM
12:00:00PM
2:00:00PM
4:00:00PM
6:00:00PM
8:00:00PM
10:00:00PM
0
1
Hawthorne Bridge
Weekend
Weekday

Annual Average Daily Bicyclists
(AADB)
4,438 4,659 4,682
0
1,000
2,000
3,000
4,000
5,000
2012 2013 2014
AADB

Compare Bicycle & Pedestrian Data
• 12% of non-motorized traffic is walking at peak
hour on south sidewalk

AADB Comparison
0
1,000
2,000
3,000
4,000
5,000
2012 2013 2014 2015
Hawthorne
Key Bridge
Source: David Patton

Phase II: Add “Explore Data” Page
Home Page
Password protectedOpen to anyone
About Page
Add Data Page
(User Dashboard)
Create
Segment Area
Create Facility
Create Flow
Create
Detector
Create Flow-
Detector
Upload Data
Explore Data
Page

Questions?
To get involved contact
Krista Nordback
Nordback@pdx.edu
503-725-2897
bp.its.pdx.edu
bp-demo.its.pdx.edu

Hawthorne Bridge
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
2012 2013 2014 2015
Hawthorne

Change over time
-
1,000
2,000
3,000
4,000
5,000
6,000
7,000
Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec
Hawthorne Bridge Bicycle Counts
2012 2013 2014 2015
20% decrease in Hawthorne bicycle counts after Tilikum Crossing opened (700/day)

Why measure walking & biking?
•Funding & policy decisions
•To show change over time
•Facility design
•Planning (short-term, long-term, regional…)
•Economic impact
•Public health
•Safety

What good are counts?
• Funding!
• Facility Level
• Change Over Time
• Planning and Design
• Safety Analysis
• Validate Regional Models
• Prioritize Projects
• Bicycle Miles Traveled (BMT)
• Signal Timing

Signal Timing
Vehicle Delay
Kothuri, S. M., Reynolds, T., Monsere, C. M., & Koonce, P. (2013). Testing
Strategies to Reduce Pedestrian Delay at Signalized Intersections. A Pilot Study in
Portland, OR. Paper presented at the 92nd Annual Meeting of the
Transportation Research Board, Washington, D.C.
Pedestrian

What?
People actually
bike here?
Yes! 200 per day

What?
People actually walk here?

Conclusions
• Data sharing makes the most of the data we have
• Bicycle and pedestrian count data are complex
• Designed for compatibility
• Connecting a “Detector” with a “Flow” via a table adds
versatility to the schema
– Allows archive to handle mobile counters
– Allows multiple counts of the same flow/time (as for
validation counts)
• Minimizing data in count data table
– Saves memory
– Improves performance/efficiency

Next Steps in Phase I
• User data input interface
• Automated upload
• User data output interface
• Basic QA/QC

Phase II and beyond
• Future Phases (unfunded)
– Enhanced QA/QC
– Analysis tools
• Summary Statistics
• AADT from short duration counts
• Integrating with weather data

Use Case – Mobile counters
The same detector can be associated
with multiple facilities and flows (at
different times).

Use Case
Validation Counts – Manual counts checking automated
counter
• Multiple counts of the same flow at the same time with
different “detectors”
DETECTOR 1
DETECTOR 2

Schema Elements
• Segment Area
• Facility
• Flow
• Detector
• Count Descriptor
• Count Data

Segment Area
Segment Area
A segment area is a stretch of transportation right-of-way over which the
volume of non-motorized traffic is not expected to substantially change.

Example Flow
Measured Flow:
Westbound Bicyclists
Unmeasured Flow:
East- and westbound
Pedestrians

Schema
Seg-
ment
Area
Detectors
Facilities
Flows
Count
Descriptor
Count Data
CD Id
Start Time
Measure Period
Volume

Highlights
• Online database bp.its.pdx.edu
• Upload non-motorized counts
• Download online
• 4.8 million records loaded for 5 states
• Demo-site
• API

Records by Year
0
500,000
1,000,000
1,500,000
2,000,000
2,500,000
3,000,000
Data in Bike-Ped Portal by Year
Number of Records Total Volume

Count Data
Sources
Bike-Ped Portal
Database
Bike-Ped Portal Web Site
Semi-
automated
ftp uploads
Data
Uploaded
via Web
interface
Raw
data
Validated
data
Meta-data
Email with
approval link
(automated
uploads)
Bike-Ped Portal
System Architecture
Visual
Validation
Interface
Data Upload
Interface
Data
Upload Script
Rejected
Data
Automated
QA/QC Checks

Data Upload Summary
Krista Nordback

Comparing Automated to Manual
< 5% error

Segment Areas
Colorado
3%
Oregon
81%
Texas
1%
Virginia
7%
Washington
8%
SEGMENT AREAS:
343 TOTAL

Modes of Travel
Bicycle
65%
Equestrian
0%
Off-Road
Motor Vehicles
0%
Other
9%
Pedestrian
26%
MODES

Arlington, VA
Bike Arlington
www.bikearlington.com

http://www.bikearling
ton.com/pages/biking-
in-arlington/counting-
bikes-to-plan-for-
bikes/counter-
dashboard/
Weather
Sunday Saturday

http://www.bikearling
ton.com/pages/biking-
in-arlington/counting-
bikes-to-plan-for-
bikes/counter-
dashboard/
Sunday Saturday

http://www.bikearlington.
com/pages/biking-in-
arlington/counting-bikes-
to-plan-for-bikes/counter-
dashboard/
Wednesday
7pmnoon7am

CDOT 2011 Bike-Ped Interface
(AVID)

Example
Short Duration Sites: 200
Permanent Sites: 20
Count Records: 30,000
Peak Hour: 1,500
Peak Day: 15,000
Max AADB: 3,000
Selected Year: 2014
Selected County: Multnomah
Short Duration Sites: 200
Permanent Sites: 20
Count Records: 30,000
Peak Hour: 1,500
Peak Day: 15,000
Max AADB: 3,000
0
1,000
2,000
3,000
4,000
5,000
2012 2013 2014
AADB
High Volume Site (Peak Hour > 60)
High Volume Site (Peak Hour 20 to 60)
Low Volume Site (Peak Hour < 20)
Selected Metric: AADB

Questions for TAC on
“Explore Data Page”
• What information should be displayed
immediately?
• How should counts be aggregated?
• For which detectors should data be displayed?
• All detectors in the archive?
• Just the detectors with the most recent data?
• Map
• Should there be a map?
• What should be on the map?

Phase II
FHWA funded
• AADT estimation for new sites
NITC funded
• Basic Data Quality
• Quality notes from data provider
• Improved data warning flags
• Communication with data
provider
• Explore Data Page
• Usability
• edit metadata
• Maintenance
• Data Wrangling
Need funding for
• Manual data user
interface
• Input interface
improvements:
• Draw segment area as
polygon
• Intersection schema
design
• Intersection schema
changes
• QA/QC enhancements
• TMG format
output/input

What is our purpose?
• Calculate monthly or annual average (AADT, etc)
• Chart trends over time
• Made data accessible
• Promote consistent and reliable bike/ped data
• Prioritizing ped/bike projects
• Other Purposes from TAC:
• Counter Maintenance
• Corridor Analysis

Use Cases
1. Calculate monthly or annual average counts
2. Chart trends over time (year over year change)
3. Make data accessible
4. Show decision-makers the data
5. Evaluate the effects of new infrastructure
6. Compare to other communities
7. Understand the impact of weather
8. Compare to NHTS/ACS data
9. Prioritize projects
10. Crash exposure measures
11. Corridor analysis

Use Cases
1. Calculate monthly or annual average counts
2. Chart trends over time (year over year change)
3. Make data accessible
4. Show decision-makers the data
5. Evaluate the effects of new infrastructure
6. Compare to other communities
7. Understand the impact of weather
8. Compare to NHTS/ACS data
9. Prioritize projects
10. Crash exposure measures
11. Corridor analysis
Addressed in Tier 1

Daily Count
0
50
100
150
200
250
300
350
400
450
500
8/25/2009 10/14/2009 12/3/2009 1/22/2010 3/13/2010 5/2/2010
Daily Count

Hourly Count
0
50
100
150
200
Total

Future
Summary
plots
Weekday Average Hourly Counts
Weekend Average Hourly Counts

Future
Summary
Plots
Average Daily Counts by Day of Week
Average Daily Count by Month

Segment Area
Facilities
FlowsDetectors
Conceptual diagram of the pieces of the schema.
Segment area is the largest rectangular region.
(Think the Hawthorne Bridge in Portland, OR)
Facilities are the smaller green rectangular regions.
(Think sidewalks or bike lanes, for example)
The lines represent flows.
The large black dots represent detectors.

Seg-
ment
Areas
Name,
State,
County,
TMG
direction,
Functional
class,
Speed
limit,
National
highway?,
Route &
number,
Observed
land use,
Start/end
date
Geometry
Detectors
Short Name
Organization
Jurisdiction
Description
Make/Model/Serial#
Automated?
Facilities
Description
Type
Side
Paved?
Width
Over or Underpass?
Sharrows?
Bike Route Signs?
Bike boulevard?
For bike lanes and cycle tracks:
-Color
-Placement of color
-Buffer
Flows
Directions
Mode
•Pedestrian
•Bicycle
•Equestrian
•Off-road Vehicles
•Motor Vehicle
•Other
Data
CD Id
Start Time
Measure Period
Volume
Count
Descriptors
Detector ID, Flow
ID, Start/end dates,
Location
(latitude/longitude)
Schema

• A Segment Area represents a segment of roadway with multiple facilities
(lanes) and multiple modes of travel
Segment Area ER Diagram

Facilities Example
Zooming in on a cross-section of the Hawthorne Bridge Segment area, we see the
individual facilities that make it up. In our context, a facility is a demarcated portion of
the roadway, such as a traffic lane, bike lane or sidewalk. On the Hawthorne Bridge, the
facilities include a north side shared use path, two westbound traffic lanes, two east
bound traffic lanes, and a south side shared use path.
Google Maps

• A Facility represents a facility along which people travel
• People may use multiple modes of travel along a facility (bikes, walking, horses)
• An example of a facility is a sidewalk or a bike lane
• Path_type can be : 'roadway’, 'path/trail’, 'sidewalk’, 'cycle track',
'bike lane’, 'general activity count'
ER Diagram

Flows and Detector Example
Measured Flow:
Eastbound Bicyclists
Unmeasured Flow:
East- and westbound
Pedestrians

•A Flow represents a flow of vehicles –
might be bikes, or peds or horses …
•A Flow may be bi-directional or single-
dirctional
ER Diagram

• Detectors represents a physical detector
– no location information
• Detectorid is key
• Serial_num also identifies detector
• Handles mobile detectors well
ER Diagram

count_descriptor
• Count Descriptor connects detectors and
flows
• A Count Descriptor represents an
installation of a detector at a particular
location – note location information
(geom)
• Designed to handle permanently
installed detectors and mobile detectors
count_descriptor_id
ER Diagram

• The Data
• Simple and sweet
• No unnecessary attributes!
count_descriptor_id INTEGER
ER Diagram
count_descriptor
count_descriptor_id

ER Diagram
count_descriptor
count_descriptor_id
count_descriptor_id

Top three lines are for REFERENCE ONLY
Document Name
Header Row – must read as shown
The “Date-Time” is the start time for the time
period during which count occurred. Date-time
must be in the following format:
YYYY-MM-DD HH:MM:SS
Duration is the length of the time period
during which counts occurred. Duration is
a time field in the following format:
HH:MM:SS
The count is the number of road users of
the type defined as the traffic “flow”
which are counted by the “detector”
during the given “duration” after the
given “Date-Time.”
Bike-Ped Portal Data Format

Reference
Lines
The first 3 lines of the upload file can be any text
you choose:
• There is a limit of 1024 characters per line.
• Only use Column A
• These can be any text. It will be saved in the
archive with the record of the upload.
• These lines are for REFERENCE ONLY
NOTE: The count will be linked to the Count
Descriptor selected during the web upload
process. The Reference lines are NOT used to
link the count to a location or detector

Ensure only 1 flow per CSV file

Save File as CSV
If you are editing the file in Excel, save it as a CSV by
1. File/Save As
2. Click “Yes” in the dialog below.

CSV file opened in Excel CSV file opened in Notepad

Document
Name
Rules for document names (aka file names):
• The file name must not contain only letters,
numbers and underscores.
• No spaces or special characters ($,/,-, ^…).
• The file must end in *.csv.
• Don’t use names longer than 200 characters.
Other than that, the name is completely up to
the person supplying the data.
We suggest that documents be labeled with
some indication of what detector/station it is
and some indication of the facility and flow of
traffic. For example, “Hawthorne_bike_NE.csv”
indicates the location name (Hawthorne), the
traffic flow counted (bike) and the facility (N),
and direction of travel (E).

Header Row
4th row is the header row which must read as
shown:
• Date-Time,Duration,Count
NOTE: The upload script must find these rows in
order to properly upload the data. Do not
include any spaces.

Date-Time
Column
Below the Header Row, each row of the “Date-
Time” column represent the START TIME of the
count.
The Date-Time column must be in the following
format:
YYYY-MM-DD HH:MM:SS

Converting to the Date-Time
Column formatTo convert your Date-Time column to the correct format:
1) Select the relevant cells
2) Click “Home” > “Number” > “Custom” as shown below
3) Copy and Paste “YYYY-MM-DD HH:MM:SS” (without quotes) into the box below the work
“Type:”
4) Select “OK”

Converting to the Date-Time
Column format
Wrong Format Correct Format

Computing Duration
Duration is the length of the time period during which counts occurred. For example, if 25
cyclists were counted between 5:30 PM and 6:00 PM, the duration would be 00:30:00.
If you only have start times counts in your file, duration can be calculated in Excel by
subtracting the Date-Time in the following row from the Date time in the current row as
shown below.

Converting to the Duration Column
formatTo convert your Duration column to the correct format:
1) Select the relevant cells
2) Click “Home” > “Number” > “Custom” as shown below
3) Copy and Paste “HH:MM:SS” (without quotes) into the box below the work “Type:”
4) Select “OK”

Save File as CSV
If you are editing the file in Excel, save it as a CSV by
1. File/Save As

Segment Area Example
Google Maps

Facilities Example
Google Maps

Metadata Input User
Interface
Live Demo

ATS-16: Making Data Count, Krista Nordback

More Related Content

Viewers also liked (7)

Similar to ATS-16: Making Data Count, Krista Nordback (20)

More from BTAOregon (20)

Recently uploaded (20)

ATS-16: Making Data Count, Krista Nordback

Editor's Notes