Using Parcel Level Data for an ActivityBased Tour Model

1
Using Parcel Level Data for an Activity-Based
Tour Model

• TRB Transportation Planning Application
  Conference
• May 8, 2007

2
Background for Modeling

• Long Range Land Use Blueprint
• 4 Ds emphasis
• Regionally adopted
• In process of developing first Blueprint
  transportation plan
• Place3s Land Use Scenario/Analysis Tool
• Parcel level data
• Place type

3
Background for Modeling (contd)

• Limitations of zone-based model
• Many 4 Ds factors missed by zone aggregation
• Developed SACSIM (Activity-Based Tour Model)
• Familiar model (similar to SF, others)
• Based on parcel-level land use data
• Motorized Networks still TAZ-based for assignment
• Skims combine TAZ skims and direct parcel/point
  proximity measures

4
Types of Parcel/Point Data Files

• Place3s Parcel Files
• Place Type, Acres
• Dwellings
• Jobs
• Schools
• K12 (all types)
• College/University
• 4 year colleges
• Community colleges
• Paid Off-Street Parking
• Spaces
• / day, / hour

5
Types of Parcel/Point Data Files (contd)

• Street Pattern
• Intersection points by type
• Types 1, 3, 4 legs / node
• Transit stations/stops
• LRT, rail stations
• Fixed route bus stops
• Park-and-ride facilities

6
Parcel/Point Data Formulations

• Point values
• of dwellings, jobs, school enrollments, etc. at
  the parcel/point
• Buffered point values
• of dwellings, jobs, etc. within ¼ or ½ mile of
  parcel

7
Strategies for Developing Datasets

• Yield Estimation for Land Use Scenario
• Qi acrespt x yieldi
• Used for both base year (2005) and future year
  dataset
• Future year land use scenarios developed in
  Place3s
• Inventory Change
• Base year points from inventory
• Future year change from other source (Place3s,
  travel model networks, etc.)

8
Examples of InventoryChange Approach

• Street Pattern / Intersection Density
• Use actual GIS intersection points for 2005
• For future year, use Place3s comparisons between
  2005 and future year to identify change parcels
• Apply lookup rates by place type to change parcels

9
Place3sExisting Conditions
10
Place3s2030 Conditions
11
Change Parcels
12
Synthetic (Gridded)Parcels in Change Ares
13
Intersection Density Lookup Table
14
Examples of InventoryChange Approach (contd)

• Transit stops
• Use 2005 GIS inventory of stops for base year
• Identify new lines by comparing 2005 and future
  year travel model transit networks (zone-base)
• Synthesize transit stops for the new lines
• Add new stops to inventory

15
Existing Transit Stops
16
Future (Model) Transit Lines
17
Future Transit Stops
18
Scale of Data Production

• Place3s land use datasets are the basis
• Separate staff to work with local agencies,
  committees etc. to develop land use scenarios in
  Place3s
• 2 persons full-time, 4-5 part time dedicated to
  this effort
• Place3s used for many land use planning, outreach
  and public relations functions

19
Scale of Data Production (contd)

• Inventory Data
• Housing, employment, schools, transit stops
• Separate function, 5-6 staff work part time on
  this
• Episodic (updates every 2-3 years or for special
  projects

20
Scale of Data Production (contd)

• Time required to generate a full SACSIM dataset
• Starting point Complete, regional Place3s
  dataset
• Ending point Complete, runnable SACSIM dataset
• Approximate duration 1-2 weeks
• Approximate staff time 50 hours, spread between
  4 staff
• 20-30 hours running time for buffering a new file
  (single thread workstation)

21
Making It Faster

• Buffering selective re-buffering
• Up front QC of files
• Re-dos are painful
• Multi-threading / server farm (in budget next
  year)

22
Acknowledgments

• ABTM Model (Daysim) Designers, Architects
• John Bowman, Ph.D
• Mark Bradley
• Application and Shell Program Developers
• John Gibb, DKS Associates
• Parcel Data Production Process
• Steve Hossack, SACOG