U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-98-166
Date: July 1999

Guidebook on Methods to Estimate Non-Motorized Travel: Supporting Documentation

2.18 Geographic Information Systems

 

Supporting Tools and Techniques

Descriptive Criteria: What is It?

Categories:

Box with an x inside Bicycle Box with an x inside Pedestrian Box with an x inside Facility-Level Box with an x inside Area-Level

Purpose:

Geographic Information Systems (GIS) are tools for managing and analyzing data. GIS can be used to enhance bicycle and pedestrian demand forecasting and facility analysis by permitting spatially-based analysis, which might otherwise be difficult or impossible. GIS can also be used to display and communicate information relevant to bicycle and pedestrian planning.

Structure:

Broadly, GIS relate environmental and population data in a spatial framework, using location points, lines (commonly roadway links and corridors), and polygons (surface areas and analysis zones). These geographic values are linked to measurable environmental and population characteristics and analyzed by spatial relationship.

Within the field of transportation, GIS are employed as a mechanism for the physical inventory of transportation facilities, as a planning tool to relate available environmental, personal transportation and household characteristics data, as a spatial analysis tool for calculating distances and areas, as a network performance monitor, and as a vehicle for the graphic display of data and analysis in a geographic context.

Currently, non-motorized-oriented GIS applications serve a variety of functions:

  • Inventory and evaluate facilities within the non-motorized network using existing condition indexing and evaluation methods. Roadway conditions, such as pavement condition, average traffic volume, and outside-lane width, are linked to specific network links. Analysis of this data and subsequent analysis can be displayed graphically in the form of a visual map.

  • Establish spatial relationships between the location of roadway network links and their condition to off-network features (activity centers, etc.) and area population characteristics.

  • Calculate and assign probabilistic gravity values of activity centers (trip generation or attraction) to geographic areas, roadway links, and location points. Roadway links are assigned a composite score based on their proximity to trip generators and attractors. This is particularly useful in the trip assignment phase of the four-step transportation demand model.

  • Compare current conditions to future projections of travel and conditions. The effects of changes in variables of underlying models can be illustrated using a GIS. For example, a GIS could produce a series of displays or maps that show the negative impact of increased motor vehicle flows on roadway conditions for bicyclists, expressed either as a decrease in level of service or increase in condition index value.

  • Illustrate impacts and calculate costs of physical improvement scenarios in a network context. A GIS can quantify improvements in level of service, condition index, or another condition evaluation by comparing present values to projections identified in planning and modeling scenarios. A GIS can relate projected physical improvements for each link to roadway link length, estimate improvement costs per link, and calculate an aggregate improvement cost. This is particularly useful in project phasing and budgeting.

  • Assess total network performance and identify optimal routes. This use of GIS is currently limited by available technology, as it must be adapted from motor-vehicle-oriented network modeling applications.

  • Produce printed maps (e.g., maintenance scheduling, project phasing, and user maps).

  • Develop network measures (street density, connectivity, etc.) and land use measures (mix, balance) that can be related to the likelihood of walking or bicycling.

Using GIS applications requires the development of a foundation data base of geographic features within the study area, including municipal boundaries, geocoded roadway links, bodies of water, and others. This information becomes the base layer upon which subsequent layers of information and analysis will be superimposed. Additional layers can attribute values or data to established roadway links, identify and classify population groups (by income, housing value and tenure, etc.) and activity centers (by trip generation characteristic). Each layer can be manipulated individually, displayed on the computer screen in any combination or printed out to meet the needs of the analyst.

Calibration/Validation Approach:

Input model calibration can be performed within the GIS. However, initial calibration of input models (spreadsheets, etc.) may require manipulations that are more easily performed outside the GIS.

Analysis of roadway conditions requires calibration to local conditions (terrain, climate, and rider behavior) and can be achieved through public involvement (stated-preference or attitudinal surveys) and testing (actual riding and rating of segments by citizen participants).

Visual inspection of computer display and printed output provides an additional level of validation and error-correction.

Inputs/Data Needs:

GIS applications require a base level of geographic data, including study boundary lines, subdivisions of the study areas (census tract or traffic analysis zone boundary lines) roadway segments, and other features within the study area.

Geospatial transportation facility files are based on a set of standard record types: link, node, point, area, geography, and attribute. Each of three spatial feature types -- networks, point facilities, and areas -- consists of an interrelated combination of these record types defining the geometry, topology, and attributes associated with a specific transportation or background feature. Specifically:

Transportation networks are composed of four related record types: link, node, geography, and attribute. Examples of transportation networks are highways, public transportation, bikeways, railroads, and waterways.

Transportation point facilities such as airports, transit terminals, and bicycle parking facilities require only two related record types: point and attribute.

Areas are made up of three related record types: area, geography, and attribute. Features such as congressional districts, states, and national parks are examples of areas.

In addition to this base layer of data, information on roadway segment conditions and geocoded activity centers may be required parts of input models. The addition and display of recognizable features and landmarks can help orient non-technical planning staff and the public.

Potential Data Sources:

Geographic data are available from a range of sources, including the U.S. Department of Transportation's Bureau of Transportation Statistics (BTS). BTS distributes the Census Transportation Planning Package (CTPP) and the Census TIGER/line files, the most broadly used geospatial data sets.

The Census Bureau's Census Topologically Integrated Geographic Encoding and Referencing (TIGER) system automates the mapping and related geographic activities required to support the decennial census and sample survey programs of the Census Bureau starting with the 1990 decennial census. The TIGER/Line files contain data describing three major types of data:

  • Line features, including roads, railroads, hydrography, miscellaneous transportation features, and selected power lines and pipelines right of ways;

  • Landmarks, including point landmarks such as schools and churches and area landmarks such as parks and cemeteries; and

  • Polygons, including geographic entity codes for areas used to tabulate the 1990 census statistical data locations of area landmarks.

The U.S. Geological Survey (USGS) produces geospatially correct quadrangles -- digital images of an aerial photograph in which the displacement caused by both camera tilt and by terrain have been corrected. USGS quadrangles combine the image characteristics of a photograph with the geometric qualities of a map, allowing GIS users to link geospatial data to the photographic image. This real world imaging can be particularly useful in interfacing with nontechnical staff, public officials and citizens.

Most state and many metropolitan transportation agencies maintain supplementary GIS-compatible data on local features, roadway classification and lane widths, and other transportation-related features. Some local municipalities have begun gathering bicycle and pedestrian-specific feature information using global positioning system (GPS) technology. The GPS produces an accurate geospatial value for location features such as curb ramps, bicycle parking and multiple-use pathway facilities.

Computational Requirements:

GIS require specialized software applications (e.g., ARCView®, ARCInfo®, MapInfo®, and Maptitude®), geographic data base files, and spreadsheet models. Complex computation and detailed graphics displays utilized by GIS are more efficiently run on desktop personal computers with fast processors and large amounts of memory or on larger network servers or mainframe computers.

User Skill/Knowledge:

Effective use of GIS requires a relatively high degree of competence in relational data base management, aptitude for computer-based data manipulation, and a working understanding of the GIS application being used. Users must also be familiar with cartographic layout principles to produce printed output.

Assumptions:

GIS applications assume the validity of their input models (condition index, level of service, latent demand scoring, etc.). When using two or more analytical models simultaneously, it is assumed the operator has taken steps to register variables common among the models. It is further assumed that the base geographic and geocoded data are in a compatible format, and are valid in location and orientation.

Facility Design Factors:

Any number of facility condition variables can be assigned to each roadway link, such as functional classification, travel-lane widths, and pavement conditions. Analysis of these variables can yield a composite score or rating for each link. These composite scores can be compared to optimal or preferred targets (derived from planning methodology or general policy goals) to identify areas which may require facility design improvement.

Output of a GIS analysis can also be applied to facility design policies and paradigms. For example, the condition index variables developed in the Buffalo, NY, area (Beltz and Burgess, 1998) were applied with modifications to the facility planning model detailed in the FHWA report Selecting Roadway Design Treatments to Accommodate Bicycles (FHWA, 1994) to produce a draft set of recommended physical improvements. Analysis using the FHWA methodology for "Group A" (advanced or experienced) bicyclists yielded a recommended facility -- a wide curb lane 4.3 m in width. Beltz and Burgess employed a GIS to calculate improvement recommendations for each link in the Buffalo, NY, study area using this method (see "Real World Examples").

Output Types:

The primary outputs of a GIS are: (1) electronic graphic display of thematic layers of data (e.g., road network or population); (2) the printing of thematic maps of the geographic study area; (3) the calculation and assignment of values to geographic areas (e.g., census tracts or traffic analysis zones) based on population, land use, other characteristics; and (4) the calculation and assignment of values to roadway links and nodes (commonly, intersections) based on proximity to trip generator locations and populations, which may become a base for trip assignment in a classic four-step travel demand model.

Aerial photo of a neighborhood
Figure 2.18: GIS can be used to develop network measures (such as street density or connectivity) and land use measures (such as mix or balance) that can be related to the likelihood of walking or bicycling.

Real-World Examples:

Buffalo, NY - Beltz, Burgess and Landis employed a GIS as a base for roadway condition analysis using the Bicycle Level of Service (BLOS) in an examination of a 1,300-km study network (center line km) in the Buffalo, NY area for the Niagara Frontier Transportation Commission. Roadway condition data were collected and attributed to individual roadway links. Each link was evaluated using the BLOS scoring methodology, inspected for validity, and assigned a composite score. BLOS scores were applied to the LOS A through F scale (with A being most accommodating and F being total lack of accommodation), and scale values for each link were themed by color for visual inspection. Through public involvement and consultant recommendations, target levels of accommodation were designated: minimum LOS C for all links and LOS B for certain priority routes and where opportunities exist.

A secondary analysis of roadway condition data using criteria from FHWA (1994) yielded a set of draft improvement recommendations specific to each link in the study network. Thematic maps were inspected by agency transportation planners, and final recommendations were made factoring in various policy initiatives study goals and other factors (e.g., parking requirements and available lane widths).

Warwick, RI - The Warwick Bicycle Network Study (Beltz and Burgess, 1998) employed a smaller scale application of GIS methods. Trip generation estimates were calculated as a function of employment, school enrollment and total population for traffic analysis zones adjacent to the study alignment. Composite trip generation scores were then attributed to network segments within the areas of influence of trip generators. The results of this analysis were compared to the existing designated bicycle route network. Alternative route designations suggested where undesignated roadway links' potential scored higher than a parallel or adjacent designated route. The results of this sketch planning effort served as the basis for final facility improvement recommendations.

Seattle, WA - The city of Seattle has created inventories of its pedestrian facilities using GIS, including existing sidewalks and sidewalk deficiencies; locations of marked crosswalks; conditions, locations, and needs for curb ramps in neighborhood commercial areas, and locations where curb bulbs are needed. This information is being matched to elementary school walking zones (305 m around each school), neighborhood service providers (e.g., libraries), and neighborhood business districts. Locations containing all three of these land uses are top priority; those containing none are lowest priority. Using the GIS inventory, recommended walking route maps for schools have been developed for each of the 60 elementary schools in Seattle. City program managers find that providing GIS-based products (maps) to the public generates increased demand for facility improvements and adds priority to proposed projects.

Orange County, CA - Hsaio (1997) of the Orange County Transit Authority used GIS techniques to analyze pedestrian accessibility to transit in Orange County, CA, using the actual street network and population information by census tract. Among other uses, the technique can be used to estimate the impacts on transit catchment population (and potentially mode choice) of improvements to the pedestrian network.

Fort Collins, CO - The city of Fort Collins, CO, used a GIS to monitor level of service (LOS) for pedestrians using a five-point LOS criteria: directness, continuity, street crossings, visual interest and amenity, and security. Areas in the city were assigned one of four designated types: pedestrian district, activity corridor, activity center, and transit route. A separate LOS threshold was set for each area type for the following factors:

  • Directness;
  • Continuity;
  • Street crossings;
  • Visual interest and amenity;
  • Security;
  • Pedestrian district;
  • Walking to schools/parks;
  • Activity corridors and centers; and
  • Walking to/from transit.

Montgomery County, MD - The Maryland-National Capital Park and Planning Commission (M-NCPPC) used indices for bicycle and pedestrian-friendliness, similar to the Portland, OR, Pedestrian Environment Factor. Each Traffic Analysis Zone (TAZ) is assigned a value according to sidewalk quality (six point scale), land use mixing (four point scale), building orientation, transit-stop conditions, and bike infrastructure. The M-NCPPC maintains a county-wide inventory of sidewalks and transit stops using a GIS.

Ames, IA - Mescher and Souleyrette (1996) used a GIS to assign bicycle condition index (BCI) values to the city street network of Ames, Iowa. The purpose of the case study was to: (1) identify optimal bicycle routes; and (2) compare them to existing and proposed bicycle route locations. The Ames BCI was developed utilizing the Delphi technique, using the Internet to coordinate expert panelist responses. The resulting BCI was used to assign penalty values to individual roadway links according to each of 14 criteria. The GIS then calculated composite scores for each roadway link included in the study.

The researchers developed an optimal route-planning tool, using a shortest-path FORTRAN algorithm, that minimizes the sum of (negative) link scores between two identified nodes. Optimal route calculations between nodes were then visually inspected for validity (based on general knowledge of bicyclist route preferences), and appropriate changes in variable weighting made. The outputs of the optimal route calculations were then compared to existing bicycle routes. In test cases, optimal routes scored significantly better than existing routes, using the identified criteria.

Melbourne, Australia - Wigan, Richardson and Brunton (1998) used a GIS to investigate trip generation characteristics of an existing multiple use trail (Lower Yarra Trail - a well-connected and promoted facility) based on user surveys, adjacent population demographics, and connectivity to these residential areas. The results of this analysis then became the basis for predicting potential levels of use at another multiple use trail (the Maribrynong Trail, which does not benefit from equivalent access and public promotion), assuming similar conditions existed.

From user surveys, Wigan et al., developed a trip length distribution model and trip generation rates for postal code zones within the study area. These rates were compared to population densities in postal code areas at various distances from the Maribrynong Trail and calculated distances from the trail to the postal area centroids (geographic center of the polygon area). Using this sketch plan method, researchers estimated a potential 500 percent increase in use, if improvements in facility access and promotion were undertaken.

Guelph, Ontario - Aultman-Hall, Hall, and Baetz (1997) use a GIS network data base to determine the characteristics of 397 routes used by commuter cyclists in Guelph, Ontario, and to compare them to the shortest path routes between each origin and destination. The analysis provides useful insight for understanding factors affecting travel behavior such as grades, intersections, etc. The study recommends different priorities for improving conditions for existing cyclists and for attracting new cyclists to the network.

Use of GIS in Travel Behavior Research - In addition to assisting with realworld planning applications, GIS has facilitated research into factors influencing bicycle and pedestrian travel behavior.

Frank et al. (1997) developed measures of pedestrian friendliness using Census TIGER files, and related these measures to the likelihood of walking or taking transit in the Seattle, WA, region. The number of census blocks per hectare in a census tract was used as a proxy for the level of connectivity and density of the street network. Related work has been conducted at the Georgia Institute of Technology (Wineman, unpublished) to develop and test topological measures of the street network from TIGER files and relate these to pedestrian flows. These measures have been found to be effective at predicting the distribution of pedestrian flows on the street network. Other travel behavior researchers (c.f. Loutzenheiser, 1997; Kockelman, 1996) have also made extensive use of GIS in analyzing land use data and relating it to travel behavior.

Contacts/Source:

Bill Barber: Metropolitan Service District (Portland, OR)

Bruce Burgess, Peter Moe: Bicycle Federation of America (Washington, DC)

Lawrence Frank: Georgia Institute of Technology, City Planning Department (Atlanta, GA)

Shirley Hsiao: Orange County Transit Authority (Orange, CA)

Bill Jack: City of Seattle, Transportation Department (Seattle, WA)

Bruce Landis: Sprinkle Consulting Engineers (Tampa, Florida)

Phillip Mescher: Iowa Department of Transportation (Ames, IA)

Matthew Ridgeway: Fehr and Peers Associates (Lafayette, CA)

Timothy Traybold: Niagara Frontier Transportation Commission (Buffalo, NY)

Marcus Wigan: Oxford Systematics (Heidelberg, Australia)

Publications:

Aultman-Hall, Lisa. Fred L. Hall and Brian B. Baetz. Analysis of Bicycle Commuter Routes Using GIS - Implications for Bicycle Planning. Presented at Transportation Research Board Annual Meeting, Paper #970168, January 1997.

Beltz and Burgess, Draft Warwick Bicycle Transportation Plan, 1997.

Burgess et al., Draft Bicycle Master Plan for the Niagara Frontier Transportation Commission, 1998.

Frank, Lawrence, Brian Stone, and Eric Matthew, A Methodology to Measure Land Use Relationships with Travel Behavior and Vehicle Emissions (draft), Presented at Transportation Research Board Annual Meeting, Washington, DC, January 1998.

Hsaio, Shirley. Using GIS for Transit Pedestrian Access Analysis. Presented at the 1997 Transportation Research Board Annual Meeting, Paper #970157, January 1997.

Huang, Yuanlin. Selecting Bicycle Commuting Routes Using GIS. Berkeley Planning Journal 10, U.C. Berkeley, pp. 75-90, 1995.

Jack, William. Using GIS to Address Pedestrian Issues. Presented at the 1997 National Pedestrian Conference, September 1997.

Klosterman, Richard. TIGER: A Primer for Planners, Planning Advisory Service Report Number 436, American Planning Association, Chicago, Illinois, 1991.

Maptitude Overview (http://www.caliper.com), Caliper Corporation.

Mescher, Phillip J. and Reginald R. Souleyrette. Use of an Internet-Based Delphi Technique and Geographic Information System for Bicycle Facility Planning. Paper written for the 1996 Geographic Information Systems for Transportation Symposium, 1996.

Onsrud, Harlan J. and G. Rushton. Sharing Geographic Information, Rutgers University, Center for Urban Policy Research, New Brunswick New Jersey, 1995.

Wigan, Marcus, Anthony Richardson, and Paris Brunton. Simplified Estimation of Demand for Non-motorized Trails Using GIS. Presented at the 1998 Transportation Research Board Annual Meeting, Paper #981203, January 1998.

Evaluative Criteria: How Does It Work?

Performance:

GIS enhances the effectiveness of spreadsheet modeling techniques by providing analysts with a visual-map display of conditions and test forecasts. GIS also enables models to account for proximity -- a major factor in non-motorized mode choice -- and the clustering of conditions and target populations.

GIS applications depend on the validity and reliability of input models, and are limited to the forecasting capability of these other tools. GIS can be a useful reference in testing optimal route designation in trip-link assignment.

Use of Existing Resources:

GIS uses available geographic data (TIGER line files, geocoded Census data, U.S. Geological Survey data) supplemented by local feature data. However, some municipalities do not have a complete GIS-coded inventory of facilities. Some features within the study area may also need to be identified and assigned a geographic value manually.

Travel Demand Model Integration:

Network analysis tools currently available with GIS applications are not sufficiently robust to enable full integration.

Applicability to Diverse Conditions:

GIS is a (spatial) relational data base; individual inputs (i.e., roadway condition indices, latent demand scoring) can be adapted to reflect local conditions and include special factors or characteristics.

Usage in Decision-Making:

GIS is particularly useful in providing composite visual representation of fairly complex underlying model calculations. Citizens, public officials, and agency staff alike can easily understand information provided in printed maps and illustrations.

Maps are extremely useful at agency staff work sessions and public meetings, where participants can identify barriers and opportunities for improvement, and better perceive and address issues related to network development, connectivity, and priorities without having a background in GIS.

GIS can be used to develop comprehensive proposals for physical improvements, including detailed design and cost information. Lists of proposed projects, ranked by priority and including data on the impacts of improvements, can aid bicycle and pedestrian practitioners in transportation and capital improvement program development, program budgeting, and long-range planning.

Ability to Incorporate Changes:

Changes may be made within the input databases and spreadsheets, and in the thematic analysis within the GIS. However, since multiple input sources are likely to be used, changes in one input data set may require a corresponding change in one or more related data sets.

Ease-of-Use:

GIS is becoming more widely used and understood by transportation professionals. Many planners with experience with spreadsheet-based modeling are discovering advantages to supplemental use of GIS, and becoming more adept at its use. However, GIS applications require an understanding of both spatial analysis concepts and the specific software being used.

GIS software applications have not reached the same level of cross-compatibility as spreadsheets; users may experience some barriers to data-sharing among State, regional, and local agencies.

Comments:

GIS has not yet been used to its full potential to relate population and activity center characteristics (as they relate to personal choice factors and demand) to the roadway network, including non-motorized pathways (physical conditions, capacity, and route choice decision making). Nor has it been fully used to assess the performance of the network as a system of interconnected links.

Attempts at using existing network performance models for non-motorized network analysis have ended with unsatisfactory results. (e.g., Matthew Ridgeway, use of TransCAD in Arcata, Calif.)

Additionally, start-up costs of more complex GIS may be beyond the reach of some municipalities. The city of Seattle spent several million dollars in implementing a new GIS. However, simpler systems can be implemented more economically, provided geographic feature data are available.

The inventory of roadway conditions related to bicycle and pedestrian travel is not routinely gathered in detail sufficient to support subsequent analysis. Supplemental data collection may be cost-prohibitive.

 

FHWA-RD-98-166

Previous | Table of Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101