USF Libraries
USF Digital Collections

Evaluating land use methods for altering travel behavior


Material Information

Evaluating land use methods for altering travel behavior
Physical Description:
1 online resource (vi, 18 leaves) : ill. ;
Frank, James E
Thompson, Gregory Lee, 1946-
United States -- Dept. of Transportation. -- University Research Program
Florida State University -- Dept. of Urban and Regional Planning
National Urban Transit Institute (U.S.)
Florida State University, Department of Urban and Regional Planning
Place of Publication:
Tallahassee, Fla
Publication Date:


Subjects / Keywords:
Local transit -- Statistics -- Florida   ( lcsh )
Commuting -- Statistics -- Florida   ( lcsh )
Choice of transportation   ( lcsh )
Land use -- Statistical methods   ( lcsh )
Geographic information systems   ( lcsh )
Population -- Databases -- United States   ( lcsh )
bibliography   ( marcgt )
technical report   ( marcgt )
non-fiction   ( marcgt )


Includes bibliographical references (leaf 18).
Performed for the National Urban Transit Institute and sponsored by the University Research Institute Program, U.S. Dept. of Transportation under contract no.
Statement of Responsibility:
James E. Frank, Gregory L. Thompson.
General Note:
Title from e-book cover (viewed Aug. 16, 2011).
General Note:
"January 1995."
General Note:

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029151716
oclc - 746955965
usfldc doi - C01-00277
usfldc handle - c1.277
System ID:

This item is only available as the following downloads:

Full Text


EVALUATING LAND USE METHODS FOR ALTERING TRAVEL BEHAVIOR Final Report for Tasks 1a, 2, and 3 Principal Investigators James E. Frank Gregory L. Thompson January 1996 Florida State University Department of Urban and Regional Planning Tallahassee, Florida 32306-2030 (9041 644-4610 office (9041 644-6041 FAX


DISCLAIMER The contents of this report reflect the views of the authors, who are res ponsible for the facts and the accuracy of the Information presented herein This document is disseminated under the sponsorship of the Department of Transportation, University Research Institute Program in the interest of information exchange The U S Government assumes no liability for the contends or use thereof.


'JICtiiOCAL RIPORT STAHDAAD 1m.E PAGE """'No. NUTI93FSU4.1 2. ACQAslon No. 3 Recipient'$ Cmloo No. 4. Tl.,. ""' Sutldi i O 6. Ropcwt Oil$ EVALUATING LAND USE METHODS FOR ALTERING TRAVEL January 1995 BEHAVIOR: Final Report for Tasks 1a, 2, and 3 e. Perrttmi"Q OOdo Aulhot('S. 8. Porformlroo Aopott No. James E. Frank and Gregory L. Thompson 9. Perfotmtlg Org&nlta.IOA fohme eM Address 1 0. Wort Ho. National Urban Transit lnstJtute Department of Urban and Regional Planning, Florida State University Tallahassee, Florida 32306 t t eonvet or OrIIC Ho, DTRS 93-G-0019 12. SpomOtln{l AQtr

Table of Co n ten t s T abl e of Cont ents 0 0 0 0 0 List o f Figures . . . . . . . . . . . . . . . . . . . . . . . . . i i i Preface 0 iv Acknowl edgments . . . . . . . . . . . . . v Abstract vii Int r oduction ...... .......... . . . .... ....... .... . .... ...... 1 T ask One/A ..... . ...... .... ......... ...................... .... 2 Purpose of the Tas k . . . . . . . . . . . . . . 2 Results . . . . . . . . . . . . . . . . . . . . . . . . 3 Task Two . . . . . . . . . . . . . . . . . . . . . . . . . 1 0 Purpose of the Task . . . . . . . . . . 1 0 Acqu i s i tion and D ecoding of PUMS . . . . . . . . . . . . . . 1 0 T ask T h ree . . . . . . . . . . . . . . . . . . . . . . . . . . 1 0 Fortu itous Augmentation of Staff Resou r ces at No Cost to the P roject 11 Reconnai s sance of Florida G I S Systems . . . . . . . . . . 1 1 Criter i a for Selection of F ield Sit e . . . . . . . . . 12 Sel ection of Test Site . . . . . . . . . . . . . 14 Decoding G I S L and Use Tape . . . . . . . . . . . . . 1 4 Acquisiti on of Census STF 3 Dat a . . . . . . . . . . 15 Ex t ract ing S T F3 Data . . . . . . . . . . . . . . . 1 5 Aggreg a t ing L and Use Data to Cens u s Tract Level . . . . 15 Merging STF3 and Lan d Use Data . . . . . . . . . . . . . . 16 Dem onstration o f F easibil ity of Merging Census Tr ip Data and GIS Land Use Data .......................................... ..... 16 Limita t ions o f Census STF3 Tri p Mak in g Dat a . . . . . . 1 7 REFERE.NCES . . . . . . . . . . . . . . . . . . . . . . . 1 8


List of Figures Figure 1 Trends in Non-SOV Use: Work Trips in Larger SMAs Entire SMA ...... Figure 2. Trends in Non -SOV Use: Work Trips in SMAs Central Cities ......... Figure 3. Trends in Non-SOV Use: Work Trips in SMAs-Suburbs ............. Figure 4 Trends in Transit Work Trips: F lor ida Urban Places ............... .. Figure 5. Trends in Transit Work Trips: F lor ida SMAs 0 0 ii 4 5 6 8 9


Preface This report presents the results of Tasks 1 a, 2, and 3 of NUTI93FSU4, Evaluating Land Use Methods for Altering Travel Behavior. Results of Task 1 b are found in another report, NUTI93FSU4.2. NUTI93FSU4 was conceived originally as a five-year project in which we would contribute to the debate on the degree to which land use variables can influence travel behavior particularly in regard to transit usage. We defined Task 1 as an update of earlier work that examined changing journey to work travel patterns in Flor ida based on the 1970 and 1980 census. The schedule called for the completion of this work during the first year of the study. Task 2 called for the use of the nation-wide 1g90 PUMS to analyze the degree which state-level policy variab les influence modal split. Task 3 called for a case study in Florida in which we would merge county assessor data tapes defining land uses as the parcel level with census STF3 data describing socio-economic characteristics of census t r acts as well as jo urney to work travel behavior The schedule called for Tasks 2 and 3 to be initi ated during year 1 but not completed. Approva l was given for year one of the study in May 1993. At about the same time the Center for Urban Transportation Research (1994) produced the first draft of its own report comparing 1990 STF3 journey to work data with that from the 1980 and 1970 censuses. Their report obviated the need for our Task 1. We then asked and received approval to substitute in Task 1 a project in which we wanted to explain transit r i dership between pairs of census tracts in Sacramento by the quality of transit service between the census tracts as well as variables describing the census tracts. We hoped to contribute to the debate by controlling for transit connectivity from each zone in the analysis to d estinations of importance to transit users. The literature on the impact of such variab l es on transit behav ior d i d not control for transit service level adequately in our opi nion The Sacramento study was approved as Task 1 b, whil e the original Task 1 was renamed Task 1 a. Ill


Acknowledgments This project is made possible through a grant from the U.S. Department of Transportation, University Research Institute Program. Their support is gratefully acknowledged. The project team for this study i ncludes: Mr. David Schmitt, Research Assistant, Department of Urban and Regional Planning F lo rida State University Mr. Chris Gray, Research Assistant, Department of Urban and Regional Planning, Florida State University. Both Messrs. Schmitt and Gray are second year master's students in the Transportat ion Planning Specialty in the Department of Urban and Regional Planning Mr. Schmitt, funded by the project, collected census STF dataand wrote computer programs to merge large data sets. The largest of these were the Orange County, Flor id a land use files which he merged with census STF3 files at the census tract level. He created another program to extract desired variables from the merged data base for analysis and he ran trial regressions of journey to work variables against socio-economic and land use variables. He a l so wrote programs that created the data bases for the simulations in the first two drafts of the Sacramento study and collected the data for the analyses of variables explaining transit trip gene r ation potentials. Finally he wrote a program that is capable of uploading the entire one percent PUMS for the nation onto the Florida State University Garnet platfonn and has abstracted a trial set of variables from it for analysis. Currently he is working with Mr. Gray on assembling data sets (inc lu ding setting up Tranplan on departmental computers and running Orlando road and transit networks supplied by KPMG Peat Marwick) for the second year project. Mr. Gray already was funded by a Florida State University Fellowship but volunteered h i s time to the project. He succeeded i n clean in g up the Orange County land use file, which consisted or several million parcels level records, and consolidating it into a census-tract based file that could be merged with STF3 files. He a l so conceived and produced all of the maps for the Sacramento project. Currently, he is designing a research study as his master's thesis that will correct some of the noted shortcomings of the Sac r amento study. He and M r Schmitt are beginning to assemble data and carry out this work as part of the second year grant. We also appreciate work done for Task 1 a and the Sacramento project by departmental research assistant Mr. Chris King. Mr. King was a first year master's student funded by a iv


departmental fellowship. He collected, tabulated, and graphed all of the U.S. census STF3 data presented in Task 1 a of this report, and he collected the first set of 1990 census data for the Sacramento project. Finally, we appreciate the work of Mr. Kevin Denny, a master's student in the Department of Urban and Regional Planning who volunteered his time to the task of cleaning up the Orange County la nd use file. Dr. Mark Ellis, Associate Professor, Department of Geography, University of California, los Angeles, also provided the project invaluable assistance in introducing the project team to the literature on Poisson regression and actually performing the first runs on the initia l Sacramento data set. He subsequently has reviewed and commented on the results of other runs Numerous persons have commented on drafts of our work. We would li k e to thank Tom Matoff, Wade White, Alan Horowitz, Susan Handy, and three anonymous reviewers from the Public Transportation Planning and Development Committee of the Transportation Board. v


Abstract This document reports on the first year progress of what was initially conceived of as a multi-year study intended to control for socio-economic, land use, and transit level of service variables in analyzing whether markets exist for transit in suburban environments. Task 1 a summarizes transit trends in Florida. Task 1 b is a trial effort based on Sacramento data to bridge gaps between the urban structure-oriented and the transit level of-service-oriented studies by controlling for both sets of variables. Task 2 involves developing a program to abstract variables from the U.S. Census Public Use Micro Sample for a later analysis of policy variables on transit usage. Task 3 carries out a demonstration of the feasibility of merging of assessor land use files with U S. Census Summary Tape Files 3 for Orange County at the census tract level of detail. Task 1 a reveals general decline i n Florida non-SOV use that mirrors national experience. The decline is not uniform, however, and the trends of transit usage in some urban areas have run counter to the average, suggesting that policy variables might make a difference. In Task 1 b (reported separately) we successfully developed a method to use transit-system generated data to reveal what suburban tracts have potentia l travel demand and which do not, and we analyzed what demographic variables contributed to demand. I n Task 2 we successfully abstracted several trial variab les f ro m the 1990 PUMS data for later analysis. I n Task 3 we successfully merged the assessor's files for Orange County, Florida with the U.S. Census S TF3 files at the census tract level. We then regressed 1990 travel behavior variables against both demographic and land use variables, but we lacked transit connectivity variables for the census tracts. What we have leamed in Tasks 1 b and 3 we will use in a study of the impact of land use, demographic, and transit connectivity variables on transit markets in Orange County, Aorida. This study will be conducted at the traffic analysis level, considerably more disaggregated than census tracts. vi


EVALUATING LAND USE MET HODS FOR ALTERING TRAVEL BEHAVIOR Fin a l Report for Tasks 1a 2 and 3 Introduction Th e salient characteristic of trave l in a decentralizing society Is ever more dispersed trip m ak ing. Users of the census describe this phenomenon in terms of overwhelming preponderance of suburb to-suburb work trips com pared to two other categories : travel within cen t ra l areas and travel from suburb s to centra l areas !Pisarski, 1987; Ros setti a nd 1993; Center for Urban Transportation Research, 1994}. Policy analysts hav e commen ted on such decentralization and its coun t erpart, the decline of central b us i ness districts (CBDs}, since at least the 1960s ( M eyer, Kaln a nd Wohl, 1966}. Policy analysts disagree on how transit systems should respond to such trends. One g roup argues that the suburbt o.CBD and the traditional inner cit y markets, despite their relative decl ines. remain t he only markets where transit can maintain model share (U.S. Department of Transportation, 1987; Jones 1985; Pisar ski, 19921 Others argue that transit should attempt to serve the non-traditional suburb tosuburb markets, because this is whe r e the bulk o f t r ave l demand is. Reasons for transit failure i n these areas are more r elated to poorly thought-out route structures that fail to allow for many -tomany travel than to the nature of suburban physical and social structur e (Thompson, 1977; Washington and Stoke s, 1 988}. The debate over appropriate transit markets has not been resolved, because previous stud ies have not controlled for variables necessary to resolve it. Studies c l a i ming that urban form or variables are o f paramount importance in expla ini ng transit success or failure do not control adequately for transit level of se rvice. Studies conc l uding that transit lev el of service is o f paramount i mportance do not control f or urban form or socio-economic variables. This document reports on the first year prog r ess of a two-year study intended to control for socio-economic, land use and transit level of service variables in analyzing whether ma rkets exist for transit in suburban environments. Task 1 a summarizes transit trends in Florida. Task 2 involves develop ing a program to abstract variables from the U.S Censu s Public Use Micro Sample for a later anal y sis of policy variables on transit usage. Task 3 c arries out a demonstration of the feasibility of merging of assessor land use files with U.S. Census Summary Tape Files 3 for Orange County at the census tract l evel of detail Our study contains another task, which because of its comp l exity is reported separately. Task 1 b is an attempt to bridge gaps between the urban structu re-orie nted 1


and the transit level -of -servi ce-oriented studies by controll ing for both sets of variables. It does so by analyzing transit patronage in Sacramento between pairs of census tracts as a function of how well transit and c:!tli1nect the two tracts, population and job densities of the two tracts, income and transit dependence characteristics of the two tracts, and their design features. Design is measured with census data describing the percentage of households before 1940. The Sacramento study as well as Task 3 of this study are precursors to a more detailed study now und erway based on Orange County, Florida. The Orange County study will use the Sacramento method with a more disagg r egated zone system (traffic analysis zones, or TAZsl. Travel behavior will focus on work trips rather than an aggregation of all travel as in the Sacramento study. By merging STF3 files and assessor's files at the T AZ level with the method that we now have proven to be feasible in Task 3, we will explain transit work trip traffic as a function of demographic, land use, and transit connectivity variables. The following pages report on the status of the work required to fulfill Tasks 1 a, 2, and 3. Task One/A f.urpose of the Task In 1991 the Center for the Study of Population at Florida State University received a contract with the F lo rida Department of Transportation to examine the implications of changing demographic trends on transport ation policy. As part of the contract, the center analyzed reports from U.S. Bureau of the Census Summary Tape Files 3 and 4 (STF3 a nd STF4l for 1970 and 1980 to determine changing trends in journey to work patterns in relation to those of the United States. Task 1 a requires an update of the trends pertaini ng to Florida using tabulations from the 1990 census To a large extent Task 1 a was made redundant with the publication of a publication from the Center for Urban Transportation Researc h (1993), but because we already had completed substantial work on the task, we report the results here. We originally wanted to update the 1970 to 1980 trends to serve as a background for our analyses in Tasks 1 b, 2 and 3, which are intended to develop an understanding of what if any policy variables migh t increase nonSOV use. We anticipated that the compilation of trends in non-single occupant (SOV) use to 1990 would show a continuation of declines noted between 1970 and 1980, as suggested by national trends through 1990 (Pisarsk i, 1992; Rossetti and Eversole, 1993). We felt a need to confirm this as a first step in investigating the feasibility of reversing trends. 2


Results The resu lts of the updates are presented in Figures 1 throug h 5. Fig ures 1 through 3 show trends in nonSOV use in the Florida SMAs that existed in 1970, differentiating between parts of the SMAs. Figure 1 shows trends for the entire SMAs, Figure 2 does so for the central city parts of the SMAs, while Figure 3 does so for the rema inder of the SMAs. As for the United States as a whole, all forms of nonSOV work trips declined in Florida between 1970 through 1980, and the declines conti nued through 1990 for all modes except one, which was working at home. Trends are similar for all parts of the SMAs; what makes the central city parts of the SMAs different from the suburban parts is the significantly greater use of non SOV modes in 1970. What is interesting is that the absolute magnitude of decline in transit use was greatest in the central areas over the twenty year period. Car pooling a lso fell faster in the central areas between 1970 and 1980, but in declined substantially in the suburbs as well as the central areas between 1980 and 1990. Overall carpooling attracted by far the largest share of non SOV use in 1970, but it decreased the most by 1990, with the declines since 1980 particularly pronounced. 3


Figure 1. Trends in Non-SOV Use: Work trips ih L arger SMAs Entire SMA (follows this page} 4


14% ,.. ,., (/) a. ;:: 12% I.e 10% 0 s 8% 0 6% c: 4% .... Q) a. 20' 10 .... Oo>l r> 70 r; ... :r,: : ; : ;.:; : . ::.;:;: Figure 1. Trends in Non-SOV Use Work Tr ips in Laraer.,SMAs Entire SMA Year . . :: . . : liKe., other work at home car passenger transit


Figure 2. Trends i n Non-SOV Use: Work frips in SMAs Central Cities (figure follows this page) 5


II) 14% c. ;:: 12% 1-..lo:: 1 0 % '""' ... 0 S: 8% lf,r' '0 6% '' ... 4% c:: 2% Q) a. 0% ' Figure 2. Trends in Non-SOV Use Work Trips in SJ\IJAs::.'LCentral Cities bike, mtr .cl., other 1980 . ; Year car passenger


Figure 3. Trends in Non-SOV Uso: Work Tr ips in SMAs Suburbs lfiguro follows this page) 6


(/) .Q.. 14% .... ..... 12% 1Qo/o 0 $ 8% . s' 2 4:: : 1 c 70 ' 2% .... Q) a. Figure 3. Trends in Non-SOV Use Work Trips in SMAs Suburbs Year taxi walk hikP.,other work at home


Figures 4 and 5 present trends In transit modal share of work trips in different Florida urban places and SMAs, respectively. Figure 4 shows a wide range in transit use in 1970, with the Miami and Miami Beach dominating transit use in the rest of the state St. Petersburg and Jacksonville trailed in a distant second place. Over the ensuing twenty years Miami Beach's transit use fell slightly f o r the first half of the period and more substantially duri ng the second half, whil e Miami's fell substantially over the first half but only slight l y over the second half. Jacksonville experienced modest decline in use between 1970 and 1980 with a catastrophic decline to 1990, while St. Petersburg experienced a severe but more moderate decline than Jacksonville. In contrast, both Ft. Lauderdale and Hollywood (both in Broward County, adjacent to Dade County to the north) exper i enced substantial gains. By 1990 Miami Beach and Miami still led in transit use by wide margins, but the next largest trans i t rid i ng hab i t occurred in Ft. Lauderdale. Jacksonville had fallen to the bottom when compared to Florida's other larger cities, while St. Petersburg had fallen into the center of the pack. Figure 5 shows much the same trends as Figure 4, but entire SMAs are portrayed. Adding the rest of Dade County to Miami and Miami Beach dilutes the more transit oriented parts of the county with lower density more auto-oriented areas, which is reflected in a more than halving of the observed modal share. Dade County still exhibited the largest riding habit I n 1970, however, and whil e it declined modestly from 1970 to 1980, it declined only slightly lower by 1990. The dramatic decline of Jacksonville stands out more in Figure 5. On the other hand, Broward County shows a modest increase. The widely different experience of trends portrayed in F i gures 4 and 5 suggests that policy might infl uence travel behavior, at least at the margin The purpose of Tasks 1 b (reported separate l y), 2, and 3 is to examine policy variables that might i nfluence 7


Figure 4. Trends in Transit Work Trips: Florida Urban Places (figure follows this page) 8


ro C/) ro -c 0 ::2: Figure 4 Trends in Transit Work Trips Florida Urban Places 20% 15% 10 % 5% 0% Urban Place 19801970 1990


Figure 5. Trends in Transit Work Trips: Flor ida SMAs (figure follows this page) 9


-. LO ::J Cl LL '#. ?{!. #-;ft. ?ft. 0 (X) <0 -.t N 0 ...... aJB4S IBPOIJ\J 0 m m ......


transit and other non-SOV use. Task Two Purpose of the Task The purpose of this task was to acquire and decode the 1990 Census Public Use Microdata Set (PUMS) and to extract a subset of the data representing the transportation variables as well as the socio-economic variables likel y to i nfluence trip making behavior at the individual traveler's level. Acquisition and Decoding of PUMS The PUMS dataset was acquired for the entire nation, a set of seven cdroms. The data was recorded in fixed length ASCII records that were decoded i n straightforward, if tedious, fashion. The data contained one alignment problem that presented a challenge. About 40 percent of the records in PUMS had an extra digit recorded in one field that the other 60 percent of the records lacked This presented problems of decoding unti l the nature of the problem was identified Then the problem was solved by a simple program that tested for the extra character upon initial read i ng of each record. Upon detection of the extra character. tha program simply branched to a r ead format that took the extra character Into account. The extraction of the subset of variables from the PUMS cdroms was time consuming in terms of machine time, often taking six to ten hou r s per cd-rom on a 486 microcomputer. The extracted files were then uploaded to a mainframe for consolidation Into a single file. Th is file is approximately 50 megabytes long. Task Three Pyrnose of the T esk The basic purpose of this task was to conduct a pilot project that invo l ved the merging of census data with land use data at a geograph ically disaggregated level in order to assemble a dataset capable to exploring the land use determinants of t ran sit trip making behavior. The i dea was to use the U.S. Census Summary Tape File 3 {STF3) data to produce census tract level data on the trip-making behavior of census respondents in a 10


selected urban area, as well as to produce data describing the socio-economic characteristics of the same resp ondents. These data would be merged with land use data acquired from the Geographical Information System (GIS) of the same urban area, thereby producing a consolidated datafile that contained detailed data, at the census tract level, on transit trip making by census tract residents, the socio-economic characteristics of those residents, and the land use characteristics of each census tract. As potential explanatory variables, that datafile would also contain Information on the land use characteristics of each census tract and the socio-economic characteristics of the population resident in each census tract. Fortuitous Augmentation of Staff Resources at No Cost to tbe Project Although the origi nal project proposal anticipated the firs t year of research dealing only with the acquisition of the STF3 data and the completion of a reconnaissance of the GIS systems in Florida's urban areas, the project's resources were significantly augmented in ways not anticipated in the original proposal, thereby allowing the work on this task to proceed much more rapidly than originally envisioned. This occurred as the result of the donated services of graduate student thoroughly versed in GIS systems and wishing to affiliate with a research environment In order to enhance his research experience. Since that person, Christopher Gray, had been awarded a University Fellowship, it was possible to obtain his donated help at no cost to the project, but at substantial banefit thereto because of his experience with G IS systems. Hence it was possible to accelerate the work program on task two as reilected in the details below. Gray's appo"mtment to the project was additionally fortuitous because it came early in the project year, facilitating a great deal of early work that then could be capitalized upon by the rest of the project personnel hired several months later. Reconnaissance of Florida GIS Systems The reconnaissance of Florida GIS systems commenced by contacting several persons known to be inform ed about GlS i n Floroda and consulting the publication by the Florida Resources and Environmental Analysis Center entitled "GIS Systems in Florida Loca l Governments" (check title). Several county GIS coordinators were contacted to discuss the general status of local government GIS systems in urban areas. At this point inquiry centered on scoping the current status of GIS in Florida, especially with regard to the potential for producing machine-readable land use data files that might be merged with the census STF3 data. 11


Criteria for Selection of Field Site Simu l taneously, the project staff began to evaluate the criteria for GIS site selection that were likely to be critical to the project. Since the fundamental phenomenon of Interest in the research was transit trip behavior, the necessity to select an urban area with a reasonable-sized bus transit system became an obvious minimum. For purposes of the research, it was important to have at least a modicum of transit usage, otherwise the examination of transit trip behavior would be useless. This narrowed the scope of the search to urban areas of at least several hundred thousand people, probably restrictin g the list to Miami, Ft lauderdale, West Palm Beach, Jacksonville, Tampa, St. Petersburg, and Orlando. A second factor that became a critical criterion was the maturity of the GIS systems considered for selection. Since GIS is a relatively new information technology and since the time that it takes to get a basic system up and running can be quite long, owing to the volume of data to be entered, the mere presence os a GIS system development project says little about the operational status of that project. It became apparent that there was considerable variation in the maturity of local government GIS systems across the state, especially in their ability to produce land use data files that were complete in their territorial coverage, well documented with regard to the details of the data, and consistent in their coding. To insure territorial coverage sufficient to encompass much of an entire urban area, especially the geographical limits of the transit network, It was decided that a county-level GIS system in the central county of a moderate-sized urban area would probably suffice, if the system was mature enoug h to have coded the entire county's land uses and to have cleaned the data to the point of being able to release it without undue trepidation about its accuracy As the reconnaissance proceeded and staff thinking on project selection criteria became more focused on the important distinctions, several other considerations had to be evaluated, includin g the issue of whether the GIS system under consideration was well situated to produce usable output files for imp ortation into the research e nvironment in which the project existed. Parallel to that was the consideration, by the research project staff, of how to configure that research environment to be able to process the GIS land use dataset to be acquired. Of course the two issues are very interdependent since the dataset to be acquired must match the processing capability available i n the research proj ect. The issue of whether the GIS site could produced a usable data tape went beyond the evaluation of the status of the land use data in the GIS system, extending to issues of whether the system could undertake custom process in g in order to produce the data needed by the project. An additional evaluat ional element was the question of how long that processing would take in relation to a one-year work program on the research grant. 12


We found that virtually all GIS systems are sufficiently overtaxed with the job of getting each layer of its system entered, cleaned, atiCl debugged that the prospect of having free resources that could be employed for custom programming was effectively nil. All the systems encounte red lacked even a p r ocedure for accepting reimbursement of the costs of custom programming. It became clear at an ear l y date that the prospect of acquiring a custom dataset was out of the question, even if project reso u rces were sufficient to pay for the work, because procedures and resources in county GIS bureaus are s impl y not set up to handle programming requests from external sources The non-customized products that are typically available from GIS systems come in two types : 11 the f i les of the GIS system as they exist, formatted for daily use in GIS system process i ng and 21 ASCII files i n text format, as fil e dumps from the re l ational database system support i ng the GIS software These were evaluated for their utility in the research project. I n the first case, the acquisition of GIS system files the project would need to acquire the specialized computer hardware and software that woul d allow the reading and processing of the GIS files i n the same manner that the host commun ity's GIS system manipulates those same files This option was expl ored with some intensity, as it had the desirable aspect of insuring maximum flexibility in the manipulation of the land use data for aggregation to the census tract level prior to merging with the census STF3 data. That is, by acqu i ring the l and use and other layers of a GIS system in the form suitab l e for manipulation by GIS software, the essential advantage of the GIS technology could ba utilized by the r esearch project I n manipulating the data. Since two GIS software vendors are dominant in Florida GIS systems, those were the two that received most of the evaluative effort by the project staff. For the project to be able to use ARCINFO GIS system files it would have required, at a minimum, the acquisition of the personal computer version of the software, called PCARCINFO, together with several equipment upgrades to accommodate the special needs of G I S systems, including massive disk storage devices, l arge central processor memory expansions, l arge format plotting machines, and high resolution graphics monitors. A second option was the acquisition of mainframe ARCINFO software and the purchase of a small mainframe {mini) computer with appropriate memory, mass storage and output devices. Both of these possibiliti e s were beyond the financial resources available in the project budget. In addition, the evaluation of the option involving the acquisition of the PCARCINFO software was tempered by the discovery that the personal computer version of the software is no longer receiving corporate deve l opment effort, thereby bring i ng into quest i on its workability o n newly emerging persona l comput i ng platforms. For GEOVISION there is no version of the program short of a mainframe version, thereby l i miting project options to the acquisition of mai nframe hardware and GEOVISION software. Th i s would have i nvolved an initia l expenditure of approximately $20,000 for hardware and an unknown amount for acquisition and yearly maintenance of the softwar e. 13


Continued hardware maintenance was an additional expense evaluated. Coincidentally, in the midst of these evaluations, the GEOVISION corporation filed for bankruptcy, adding new uncertainty to the long-term viabil ity of that system. For these reasons, the costs of acqu iri ng and manipulating the formatted GIS files was considered to be much too expensive. The use of ASCII files. while not the most flexible and elegant, turned out to be very cost-effective. It was the option implemented. This Involved verifying that the county GIS system selected had progressed in the development of its GIS system to the po in t at which it had an operational land use layer in its system that could give us parcel-level data showing land use according to a fairly detailed land use coding scheme, parcel size, building floor space, building age, building stories, and census tract within which the parcel was located. Since these are standard dataitems in mature GIS systems, and since we were willing to take them as a simple extract from the GIS database without any manipulation, they could be written to a datatape in ASCII format by a simple query to the database, something achieved by the GIS staff as a routine database dump. The key tci the efficiency of this procedure was our willingness to accept the data at the parce l level the level at which the data is retained in the GIS database system, and being willing to do our own processing to aggregate the data to the census tract level Selection of T Sjte After carefully weighing the several factors in our selection criteria and undertaking the necessa r y verifications of GIS capability, it appeared that Orange County was the best choice. Decoding G I S land Us!! I!lpe The ASCII-coded, parcel-level, land use tape of Orange County consisted of approximately 240,000 records, one for each parcel in the county. The tape was not entirely clean of problems. Some records were mis-aligned and could not be read by a fixed format, Fortran-li ke format statement. A number of records had no census tract number and had to be discarded. A few records had data missing from fields other than the census tract. Dealing with all of these problems was comp li cated by the large number of records in the file, making it impossible to use, for instance, microcomputer spreadsheet programs. The final solution, arrived at through considerable trial and error consisted of reading the tape on a mainframe tape drive, downloading the entire f i le to a personal computer importing the data into a Paradox database program; using the parsing feature to extract 1 4


characters individually and to combine them into their properly align ed fields. In the end, a file of correctly aligned records was obtained, although between five and ten percent of the records lacked a census tract number and had to be discarded. The development of the final solution to the tape decoding problem was the direct result of the tenacity and ingenuity of three very valuable people: Chris Gray, the volunteer research assistant with expertise in GIS systems mentioned earlier in this report: K evin Denny, a graduate student with considerable experience in database systems who volunteered to help out on the tape problem; and Dave Schmitt, our paid research assistant who can program a computer to do just about anything and who contributed key pieces to the so lution. Acquisition of Census STF3 Data The Census STF3 files on cdrom were acquired by the project free of charge from a fellow faculty member who had previously acquired them for his research and who was willing to loan them to this project. We thank Professor Peter Doan of the Department of Urban and Regional Planning for being so gracious. STF3 Data The STF3 data were found to be clean and of high quality. They are supplied on cd rom in DBASE format and also have a census-supplied program, GO designed for extracting the data from the disks. The p rocess of extraction is tedious, because of the way that the files are organ ized on the cdrom; that is similar data items are grouped in one DBASE file for all geographic levels. But extraction of numerous data items at one leve l !census tracts in one county) requires repeated entry of numerous DBASE files Aside from the normal fam i liarization with file structures and record layout, the data extraction was accomplished in straightforward, it boring fashion. Thanks go to David Schmitt, our paid research assistant for that fine work. 8ggregetiog Land Use Data to Census Tract Level The parcel-level, land use data from Orange County's GIS system were aggregated to the census track level by summing parcel size and build i ng floor area for each land use code within each census track and by calculating the mean building age and mean number of stories for structures for each land use code within each census track. Parcel land area, building floor area, building construction date, number of stories, and land use code were 15


the five characteristics for each parcel that had beeh obtained from the GIS database. This created a 130 record file (one for each census tract) with each record having over 800 variables, since the number of land use codes was in excess of 200 and four characteristics (land, floor area, age, stories) were tabulated for each land use code Merging ST E 3 and Land Use Data The merging of the census tract land use file w ith the census STF3 file was accomplished by a C program written by David Schmitt after uploading the two files to the mainframe computer. The use of the ma i nframe was necessitated because the SPSS was anticipated as the statistical package to be employed and microcomputer vers i ons of that program had file-size lim its that would have been exceeded by the size of STF3 extract file. It should be noted that microcomputer statistical processing could have been achieved if we had been willing to aggregate land use codes into coarser subsets of the original 200 plus l and use categories. However, that would have sacrificed the flexibility of being able to tailor the aggregation process after the merging of the STF3 and l and use f i les. Alternatively, it would have necessitated reaggregation and merging if alternate combinations of land use categories became desirable. Since we had e x cellent programming skills in project personnel, it was decided to upload to the mainframe without grouping land use codes and then use SPSS variable recomputation capabilities to combine l and use categories for each statistical run as necessary. Demonstration of FeasibilitY of Merging Census Trip Data and GIS Land Use Data At this point we had a data file for Orange County conta i ning, for each census tract, detai led i nformation on t r i p making by census respondents, together with the socio economic characteristics of those trip makers, and the l and use envi ronment in which those tripmakers lived. Since major purpose of this research project task was to explore the feasibility of merging census and GIS data, it is appropriate to review the lessons learned from that test. The feasibility of acquiring and merging the two types of data to create a dataset capable of fac i litating research into the land use determinants of trip-making behavior has been amply demonstrated to be possible without insurmountable hurdles. It requires a microcomputer with cd rom capability and personnel w ith modest DBASE skills to extract the desired STF3 data from the census files It requires at l east some access to a mainframe computer and ability to program i t to do tape copying and data extracts in order to be able to read the ASCII tape of parcel-level data. But the mainframe capability can be avoided if: 11 the GIS system f r om which the data are being acquired is w i lling to custom 16


program a GIS manipulation that the televant land use data to census tracts and 2) the researcher is willing to work with data representing combined land use codes in order to reduce the number of code categories to a more manageable number. If these two things can be accomplished, then the researcher can expect a dataset with several hundred reco rds (census tracts) and perhaps one hundred fields ( say, twenty land use codes by five parcel/structure characteristics). This size is well within the data matrix limits of SPSSPC and other common microcomputer statistical packages The merging of the two census tract files can be easily accomplished in something like SPSSPC as long as the data matrix size limitation problem is avoided. For those unable to avoid the large dataset size problem or unwilling to work with land use codes that are combined into coarser categories, mainframe processing may be unavoidable. But for most analyses, the collapsing of land use categories to something in the range of 25 to 35 should permit the creation of land use variables of sufficient detail to test most hypotheses about the effect of land use on trip-making behavior. And for many scholars, mainframe access and programming are the staff of life. For those researchers unable to persuade the GIS agency to do custom programming, who therefor are forced to turn to parcel-level ASCII tapes, things will go more smoothly if that tape is requested to be produced in comma delimited format. This will facilitate the resolution of alignment problems using database programs that have the capability of importing delimited files. Limitations of STF3 Trip Making Data It is appropriate to discuss the details of the trip-making behavior data produced by the U.S. Census. Each census respondent is asked to provide information describing their journey-to-work for the previous day, including their mode of travel, their time of departure, the length of their trip, and their destination city/county. It is worthwhile keeping in mind that these data report only work-trips, and only the home-to-work trip. This excludes non-work trips as well as the work-to-home trip. 17


REFERENCES Center for Urban Transportation Research 1993 Florida Demographics and Journey to Work: A County Data Book !Tampa: Center for Urban Transportation Research, University of South Florida). Center for Urban Transportation Resea rch 1994 Demographic & Commuting Trends in Florida !Tampa: University of South Florida):33-38. Jones, David 1985 Urban Transit Policy: An Economic and Political History (Englewood Cliffs, New Jersey: Prentice-Hall). Meyer, J.R J. F Kain, and M. Wohl 1966 The Urban Transportation Problem (Cambridge, MA: Harvard University Press):9. Pisarski, Alan E. 1987 Commuting in America: A National Report on Commuting Patterns and Trends (Washington: Eno Foundation for Transportation). Pisarski, Alan E. 1992 Travel Behavior Issues in the 90s (Washington, D.C.: U S. Department of Transportation, Office of Highway Information Management). Rossetti, Michael A. and Barbara S. Eversole 1993 Journey to Work Trends in the United States and its Major Metropolitan Areas, 1960-1990 (Washington, D.C.: U.S. Department of Transportation, Office of Highway Information Management). Thompson, Gregory L 1977 "Planning Considerations for Alternative Transit Route Structures," Journal of the American Institute of Planners 43: 168-168. U.S. Department of Transportatio n 1987 The Status of the Nation:s Local Mess Transportation: Performance and Conditions (Washington: USDOT):46. Washington, Earl J. and Robert W. Stokes 1988 Planning Guidelines for Suburban Transit Services, Final Report (Washington: U.S. Department of Transportation, Urban Mass Transit Administration). 18

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200445Ia 4500
controlfield tag 001 029151716
005 20110816181722.0
006 m d
007 cr bn|||||||||
008 110816s1995 flua obt 000 0 eng d
datafield ind1 8 ind2 024
subfield code a C01-00277
HE4487.F6 (ONLINE)
1 100
Frank, James E.
0 245
Evaluating land use methods for altering travel behavior.
p Final report for tasks 1a, 2, 3
h [electronic resource] /
James E. Frank, Gregory L. Thompson.
Tallahassee, Fla. :
b Florida State University, Department of Urban and Regional Planning,
1 online resource (vi, 18 leaves) :
Title from e-book cover (viewed Aug. 16, 2011).
May 1993-Dec. 1994.
Performed for the National Urban Transit Institute and sponsored by the University Research Institute Program, U.S. Dept. of Transportation under contract no.
DTRS 93-G-0019.
"January 1995."
Includes bibliographical references (leaf 18).
Local transit
z Florida
v Statistics.
Choice of transportation.
Land use
x Statistical methods.
Geographic information systems.
United States
Thompson, Gregory Lee,
d 1946-
United States.
Dept. of Transportation.
University Research Program.
Florida State University.
Dept. of Urban and Regional Planning.
National Urban Transit Institute (U.S.)
i Print version:
Frank, James E.
t Evaluating land use methods for altering travel behavior. Final report for tasks 1a, 2, 3.
Tallahassee, Fla. : Florida State University, Department of Urban and Regional Planning; Springfield, VA : Available through the National Technical Information Service, [1995]
w (OCoLC)32914105
Center for Urban Transportation Research Publications [USF].
4 856