A Virtual travel survey:

evaluating the regional economic impact of Transit-based corridor development

The quantification in absolute values of the socio-economic consequences of given regional growth strategy would be complex and highly subjective to the beholder’s social, philosophical and political assumptions. A quantification in dollar and cents the impact of each scenario would be impossible. Fortunately, however, what we need is not an absolute value, but rather the relative value among different scenarios.
Given that the objective is to enable leaders to make knowledgeable allocation of limited resources, much easier and equally effective is an evaluation of the comparative value of different scenarios.
Great amounts of data, if controversial, are available on the personal and social costs of transportation.
Under many social and environmental parameters, a successful PLC should clearly outperform its competitor (the car-based corridor). However, its influence on the following regional factors would have to be demonstrated:

• Regional Connectivity. (Freedom to participate in desired spatial interactions).
• Regional real-estate values.
• Regional Economic Growth.
• Equity in the allocation of regional benefits/dis-benefits.

Next, I will analyze principles of travel behavior and locational choice to develop a measure to compare different transportation/land-use scenarios.

 HUMAN DECISION PROCESS, INTERACTION, ACTIVITIES AND TRANSPORTATION

All people engage daily in a variety of actions in response to a number of desires, aspirations, and moral obligations, aimed at maximizing perceived utility.  Some courses of action are chosen by thought, some by habit, some by instinct, but always with the conscious or unconscious objective of maximizing overall perceived benefits or “utility”.
All actions result from a choice by a person (or “interactor”) to engage in a two-way interaction with a combination of objects and people (or “interactees”).
All interactions happen through many senses, or communication channels, and often involve primary and secondary objects and people.
Given the difficulty of “turning off” selected bodily senses or isolate “interactees” from their contexts, all chosen interactions also produce a number of unwanted interactions. This means that all actions we initiate bring dis-utilities as well as utilities. People will chose interactions that maximize perceived utilities and minimize dis-utilities.
Most people periodically go through an evaluation processes to determine which course of action to undertake; be it getting a glass of water or buying a house. Some decisions determine instant course of action, while others set intentions for a future action or pattern of actions. Therefore, people will choose to interact with certain objects and people at a given time, place and for a given duration.
We periodically choose among a number of perceived interaction alternatives based on their utility levels.

Urban space as maximixer of positive interactions

The utility of an interaction is dependent, among other things, on the proximity of interactor and interactees, and a preferred time of the interaction to take place. The following factors should be optimized to maximize utility:

Adequate Adjacency:    “Interactor” and “Interactees” have to be within reach of each other, either through spatial adjacency or through virtual adjacency (telecommunications). A growing number of interactions are rich combinations of spatial and virtual communications that involve many senses and many “interactees”.

Adequate Timing: The level and positivity/negativity of the utility level varies with the time and duration at which it happens. Most interactions (or activities) have a time range within which the utility level remain somewhat constant, this allow for a scheduling flexibility of the “interactor”. Other interactions, like work, have specific peak times at which utility is maximized. In fact, I will have high utility only when arriving on time. This time range can consist of minutes (ex. catching a train) or days (ex. wash car) or weeks or months (ex. painting my bicycle).

Adequate connectivity and timing are obtained through timely movement of “interactor” and/or “interactees” to achieve spatial adjacency, and through telecommunications to achieve virtual adjacency.

The maximization of these utilities is the reason why people (“interactors”) move to certain places or travel to certain locations. In most cases the desire for travel is derived from the desire for a specific interaction, in rare cases it becomes both the means and the end (ex. a ride in the park).

Mostly, however, travel is a dis-utility we are prepared to sustain as a price to obtain a certain utility: it is a negative interaction we are prepared to sustain to be able to engage in a chosen activity.

 The cost of providing adjacency to “interactor” and/or the “interactees” is borne by many different parties. For example, the cost of my driving to “interact” with people and tools at work is borne in part by the following: myself, by operating the car; my employer, by providing parking; society, by breathing the fumes.

Scheduling of activities and interactions

Both consciously and unconsciously, every person revises daily, weekly, monthly tentative schedules based on places they have/want to be, things they have/want to buy, people they have/want to meet, relatives they have/want to escort. 

Since the utility of each activity varies through time in a different way, each can be defined by certain temporal characteristics:

·         Duration: the duration of a certain activity is approximated in our “schedules”. Duration of activities might be more or less foreseeable. Its actual value is affected by a combination of external events and dynamic rescheduling.

·         Priority Level & Time Range: Activities could be categorized as primary or discretional. Primary activities (work, study…) are those that have a very high utility attached to them. Our participation and/or timeliness in these activities have great impact on our utility levels.  However, this simple classification does not describe precisely how the utility level of each activity varies in time. Some activities have a “time range”: a span of time (in hours, days, weeks, months) within which I can accommodate the activity maintaining equal utility; for example, grocery store hours are the time range for my grocery shopping activity. A single curve could represent all these factors by describing the utility/dis-utility of performing the activity at different time/dates.

A regional physical growth strategy should be based on the basis of how well it accommodates these “Wish/To-Do” Lists (i.e. how easy it is to participate in desires activities) of businesses and individuals.

ECONOMIC EFFECTS

Geographic Distribution

Changes in regional transportation and/or land-use have one major economic effect from which all other effects are derived: real estate and land values.
Properties, whose connectivity and quality of their proximity have been improved, will eventually embed that benefit through higher markets value. On the contrary, properties that sustain a worsening of the environment, proximity to negative activities, barrier effect, decreased relative accessibility, will eventually embed those dis-utilities through lower market values.

Demographic Distribution & Equity Issues

People and businesses located on the affected locations will directly suffer or enjoy those changes only if they own the property or their rental agreements are somewhat “stabilized”. In fact, in the case of leased property, the changed value of affected properties will ultimately sustained by the owner through changes in demand. This change in demand would result in an increase in the eventual rental rates to match new market value.
Renters’ utility level would temporarily be altered until rental agreements are renegotiated.
Therefore, any evaluation of the equity effects of certain Transportation/Land-use strategies should analyze where connectivity and/or environmental impacts are altered, and consider the ownership levels.
Non-owners could still be affected by a given Transportation/Land-use strategy if real estate value changes in the area somehow affect the performance of their employer or their own business. For example, a new highway exit close to a store where I happen to work will make it more likely that I will maintain my position or maybe get more responsibilities in a larger firm. If, for example, I own a pizza business at a rented-location, the improved demand for my product might offset the increase in rental premium.
However, these economic effects (derived from the trickle down economic theory) are impermanent, hard to measure, and probably minimal respect to direct property effects.
For example, many improvements in bus services are justified from the equity standpoint because they provide increased accessibility to low-income neighborhoods. However, if most residents rent property, those benefits will eventually be absorbed by the rental rate. Even when residents own the property, the increase in property value is seldom proportional to the increased accessibility; in fact, improved bus service is considered as impermanent and subject to sudden and arbitrary reductions, cancellations and re-routing.

LOCATION, LOCATION, LOCATION:  EVALUATING a regional GROWTH STRATEGY THROUGH REAL ESTATE VALUES

Transportation and land-use spatial-temporal relationships determine the availability and quality of any location time/place utilities.
In fact, aside from the conditions of local and regional real estate markets, the relative quality of a location is mainly dependent on its spatial-temporal placement with respect to other activities and locations. Regional transportation infrastructure, local land-use policies and local land developments give a location its value; if none of these existed, no practical use of that land would be economically feasible. The topology and performance of transportation network, as well as the topology, intensity and socio-demographics of developments determine its relationship to other activities. These relationships give a location its market value and can be expressed by two sets of measures or “location profiles”:

·         Regional Connectivity: It describes the relationship of a location to activity opportunities. A location has easy access to certain activity opportunities and uneasy to other. Regional connectivity is always a positive attribute; however, the connectivity to different activity type has different value to different demographics. The transportation system and land-use distribution in a regional area influence the potential utility of a location by determining the dis-utility levels that local residents (or “interactors”), have to endure to come into spatial adjacency of desired activities (or “interactees”).

·         Local Proximity: It describes the relationship of a location with its immediate surroundings. It includes the effects of local connectivity, environmental impact and adjacency effects. Every location can sustain both utility and/or dis-utility from its local connectivity, proximity and adjacency to another. Therefore, local proximity is positive or negative. For example, being across from a house more beautiful than mine might increases the value of my own house; if that house were abandoned and decayed, instead, it would decrease it.   In other instances, the relationship toward another activity could bring connectivity benefits and proximity dis-benefits. For example, being close to a convenience store might be convenient for access to shopping but might decrease household safety. The transportation system and land-use distribution affect the utility levels of locations by attaching to them specific positive or negative “interactees”. For example, a new transit system might “attach” a high level of noise to surrounding locations, or a new land-use reconfiguration might give condominium views of a park. In such case, the “interactors” using the site would have difficulties avoiding undesired interactions with noise or desired interactions with a good view.

However, transportation and land-use do not directly affect the utility of interactors, but the utility of the locations where interactors choose to locate. The “interactors”, in turn, choose a location were quality of available interaction opportunities is highest to them, relatively to its price.
Given constant location costs, people and businesses will locate spatially and temporally close to desired “interactees” and locally far from undesired ones. They will locate to maximize their positive proximities and minimize their negative ones.

MEASURING REGIONAL CONNECTIVITY OF PROPOSED URBAN GROWTH STRATEGY

To evaluate the effects of proposed transportation/Land-use policies and investment planners have been using the “scenario” method: different proposed alternatives, including a “no-change”, are compared by major performance measures to assess their relative merits.
I will try to outline a proposed method for the evaluation of radical urban scenarios through a “scenario” comparison to the “no-change”. This method will be based on the comparison on an actual activity survey and a Virtual Activity Survey.
An actual activity survey will track the location, movements and activities of a number of residents of the study area. It will keep detailed and consistent record of activities engaged and purpose of activity.
The Virtual Activity Survey will have each of the surveyed residents assigned new virtual home and job locations and replicate that same week of activities in the proposed changed urban environment.
The current and proposed urban environments have to be thoroughly surveyed and inserted into a GIS database (we will discuss later data needs).
Many travel behavior surveys have shown that peoples travel behavior is characterized by complex multi-stop tours that revolve around home and one or more mandatory/primary activities (work, study, child care, gym, …). Tours are structured along four major sets of origin and destinations: Home-to-Home, Home-to-Primary, Primary-to-Home, and Primary-to-Same-Primary, Primary-to-Other-Primary. Secondary stops are decided based on habit, personal preference and/or as a trade-off between interaction utility and their distance in time (ex. I might decide to stop at a further grocery store because of its wider choice or lower prices).
Whatever the reasons are, we benefit from the opportunity to choose among different suppliers to somehow maximize utility, even if we chose a certain supplier simply to have a different experience. The large choice provided among different suppliers of similar activities is what has made automobiles such symbols of freedom.
People, however, will value this ability to choose the supplier of a service differently. (For example, while I might decide not to drive 5 more blocks to save some change on a quart of milk, I definitely will drive to the other side of town to use the services of a “trustful” lawyer. The ability to choose among milk suppliers might not do me much good, but the ability to choose among a large number of attorneys might. This means that, whatever the personal transportation costs were to reach the other side of town, the benefit of choice was higher.
Therefore, we can conclude that a Transportation/Land-use structure provides me with the benefit of choosing among different suppliers of a service. The higher the number of similar supplier I am willing to forgo, the more I am benefiting from the freedom of choice.

“Virtual activity survey”: a proposed connectivity measure for transportation/land-use scenarios

The connectivity benefits of a location are ultimately enjoyed by the resident and businesses that locate there. Each individual and business has different interaction needs, and therefore will take advantage of different connectivities.
Each person would benefit from spatial/temporal proximity and connectivity to certain interactions and be hurt by spatial/temporal proximity to other interactions. For example, the proximity to a gourmet store might be a benefit to an high-income household, and not to a low-income one. (That’s how real estate developers are able to locate their projects).
Households will choose housing and/or job location that more closely match their connectivity and proximity profile. Although these differences are among individuals, most variation can be captured by demographic classifications. Controlling for a few demographic characteristics, the accessibility profile of people is fairly similar: they will value similarly the proximity and connectivity to different categories of activities.
The data actual travel survey records each activity supplier visited by the surveyed resident. Detailed geographical information of built environment and transport performance data will enable us to determine the number of similar activity suppliers he/she has forgone to be able to choose that particular supplier. Then, we are able to estimate the exact time that would have been saved by choosing any of the closer suppliers. Therefore, we can define the actual chosen supplier in terms of how many similar supplier could have been reached quicker from the residents home or as additions primary tour. (In the case of the attorney, I might have chosen the 78^th closest attorney).
Once all these data is collected from the actual survey, and elaborated through geographical data, we can go on with the Virtual Activity Survey, replicating the sequence of activities and the amount of choosing that occurred.
A city is good to us in so far as it allows us to participate in the interactions we desire. Simulating the same activity selection and choosing behavior we’ll be able to compare how good each given urban scenario is for its residents.

A new home location and new job location will be assigned to each of the surveyed residents. A certain resident will be assigned a virtual residential location, which mostly approximate his/her actual one in price, connectivity and proximity profiles.
He/she will then be assigned a virtual location for his/her primary activities. This will be done on the base of job type and the number of forgone potential jobs that were closer to its virtual residence.
All other activity destinations are than re-assigned in the virtual city in a manner similar to job location. For example, if in the actual survey “John” stopped at the third closest “Big Box” grocery store to its work-to-home path, then our virtual survey will also choose the 3^rd closest grocery store in the new scenario.

VIRTUAL vs. ACTUAL SURVEY

Next, virtual and actual activity surveys can be used to compare the relative accessibility of each scenario. We can compare what are the relative costs of providing a specific amount of spatial/temporal freedom to participate in certain activities. Meaning, we can analyze the relative costs in time and transportation expenses that the same individual occurs in the two scenarios for the same time/place utilities.
Differences in travel times can be visualized in several different ways, to determine how each scenario would perform respect to different demographics, different activities, different time periods. Some of the measures that could be extrapolated are:

·         Weekly Total Difference by individual or group.

·         Daily Total Differences by individual or group.

·         Continuous charted weekly travel-time comparison.

·         Comparisons across demographic groups.

·         Comparisons across different activity types.

These measures can be used in the evaluation of a given urban growth strategy, because they embed changes in regional and local connectivity. These measures, supplemented by measures of proximity effects, could give a comprehensive assessment of the impact of a certain strategy.
Alternatively, actual and virtual records of individual travelers can be used as anecdotal demonstrations of the benefits of the strategy, or examples of “a day in the life” of a proposed urban plan.  Also, they could be used by developers to evaluate the marketability of a proposed transit-based development to specific demographics.

DATA NEEDS

Data on actual survey, transportation networks and land-usage must be very detailed spatially and temporally for these methods to perform well. The location of both the traveler in time and of all potential destinations should be precisely defined and inserted in a GIS database.

Three data sets are needed:

Detailed Actual Activity Survey: Current Differential-GPS and cellular phone technologies allow tracking the movement of people within urban environments. Issues of accuracy, consistency, and privacy, are rapidly being confronted and resolved. Great commercial opportunities are pushing their developments.. In fact, a new Federal law requires that all cellular phone will be locatable within 125 meters by 2001 for 911-emergency service. USDOT will soon complete the installation of special radio transmitters that will allow the location of watch size D-GPS receivers within 4 ft. Within months these technologies might be precise enough to locate the exact establishment or land parcel where the traveler is at any given moment. Also, mode, time and duration of trip, as well as time and duration of activity would be automatically recorded. In fact, mode of travel can easily be established by the analysis of geocoded traveling speeds and patterns.. Comparing the traveler position with location of all suppliers we can establish if he is in a ice cream shop or a drugstore. To complement these data, the surveyed will be asked to confirmed tracking data and provide additional data, such as: name of place visited, detailed purpose of visit, company. Remote location would greatly diminish the burden on the surveyed individual, reducing the “non-response” problem.

Land/Space Utilization: Location and attributes of all activity opportunities should be recorded for actual and proposed scenarios. Bulkier, more expensive and more precise remote location receivers are used to geocode buildings, land parcels and transportation infrastructures. Their accuracy is important to be able to match it with travelers’ data to determine “hits”, or visits. Activity suppliers are geocoded and classified in essentially-substitutable categories. Suppliers in a same category offer very similar interaction opportunities (or similar goods and services).

In reality, no two suppliers are truly substitutable because of different quantity and/or variety (I might go to McDonald for their particular kind of junk food, not only because it is just another fast-food restaurant).     We will assume, however, that these subjective preferences within essentially substitutable categories of suppliers are relatively weak (i.e. if I cannot access a McDonald, I will settle for a Burger King without much distress).

For example, grocery stores might be classified as “Big Box”, “Mum & Pa” or “gourmet” reflecting different clienteles.

Transportation Network: Network topology and performance should be defined for current and proposed scenarios.  For the ACTUAL SURVEY, Topology can easily be acquired through remote location technologies and, for the most part, is already available for US urban areas.

Network performance (or estimated speed at a given time) is, instead, might be hard to measure.

In fact, the speed of auto links changes by hour of day, by day of the week, by season, by weather conditions, and so on. The main cause of this is the fact that flow in an auto-facility is restricted by high Volume/Capacity ratios: the number of cars on a given street at a given moment (when high enough) affects the level of flow at that same moment.

For the greatest majority of links, however, flow is fairly constant because Volume/Capacity is fairly low. Therefore, their speed through time can be reasonably approximated.

Other streets are affected by regular and/or occasional congestion, and a number of flow measurements might be required at different time of day, and day of week.

Therefore, data on current auto-links performance can be acquired; its precision will depend on the number of measurement performed, which in turn is bound to a data acquisition budget.

For the VIRTUAL SURVEY, the creation of a network performance for the proposed scenario is required. It requires that we predict the effects of given land-use and transportation changes on roads flow. This has traditionally been a very unlikely endeavor because of the many dynamic involved in the Volume/Capacity ratio and Down’s “Triple Convergence” effect.

It would be impossible to predict travel times and classify locations in order of connectivity, because the flow on each transportation segment could not be predicted.

If, instead, we are considering the impacts of a strongly non-auto-based Transportation/Land-use scenario,

our ability to simulate the actual survey is greatly improved.

We are able to measure changes in regional connectivity through the speeds of separate-ROW Transit, pedestrian and bicycling. They can be predicted because their Volume/Capacity ratio rarely reaches critical highs. With these modes, speeds can be defined given transit system design, schedules and frequencies, average walking and bicycling speeds. In fact, when transit enjoys a separate ROW, traffic signal preemption or a combination of short routing and traffic prioritization, their speeds in time can precisely be ascertained (see Fig. 1).

Such transit-based strategy should have as major outcome an increase in regional connectivity, for certain areas near transit nodes, through various intermodal combinations (transit/walk, car/transit/walk, bike/transit, etc.). Also, certain road links should improve their flow because of diminished car ownership and usage. For the “Triple Convergence” effect and other reasons, I would assume these changes in car flow to be a relatively minor connectivity improvement, and therefore can be just roughly approximated.

GIS TOOLS

Given a the complete set of topological and attribute data, GIS computer software allows to precisely calculate the shortest paths between any given destination. Proximity and accessibility can be simulated. That can, in turn, be related with existing and proposed land utilization to define the spatial/temporal relationship of different locations.

This information can be combined in a number of ways to communicate the transit/pedestrian accessibility (i.e. the freedom to participate in desired activities). Exact time costs can be calculated for each potential trip, trip chain, trip tour.

All essential tools to perform our VIRTUAL ACTIVITY SURVEY are presently included in GIS software; some programming will be needed to automate many procedures.

Following is part of the spreadsheet model I have developed to automatically calculate the time needed to move through a transit/pedestrian network within GIS. Any change in the following attributes of the transit systems would automatically update the time costs associated with each link:

Figure 1: This GIS-based transit/pedestrian link time cost calculator enables to assess precisely the amount savings in traveling times brought by a high quality and frequent transit system

Selected Bibliography

·         Ben-Akiva, Moshe and Bowman L. John.  Activity Based Travel Forecasting.Activity-based Travel Forecasting Conference.Wash., DC: DOT.

·         Ben-Akiva, Moshe and Bowman L. John. The day activity Schedule approach to travel demand analysis. Wash., DC: TRB annual meeting 1998.

·        Bhat R.C., and F.S. Koppelmann. A retrospective and Prospective Survey of Time-use Research. Northwestern University. (#990830).

·         Bhat R.C., J.P. Carini and R. Misra. On modeling the Generation and Organization of Household Activity Stops. TRB Annual Meeting, 1999.

·         Cervero, Robert. The Transit Metropolis. Washington, DC: Island Press, 1998.

·         Shen D., Huang J. and Zhao Fang. Automated People Mover Applications: a worldwide review. Washington, DC. : National Urban Transit Institute.

·         Warren, Roxanne. The urban oasis. McGraw Hill, 1997

·        Miller, J. Harvey. Measuring space-time accessibility benefits within transportation networks: basic theory and computational procedures. Geographical Analysis, April 1999.

·        Zhang M., Q. Shen and J. Susssman. Strategies to improve job accessibility- a case study of Tren Urbano in San Juan Metropolitan Region. 1998.

·        Niles J. and D. Nelson. Measuring the success of Transit-oriented Development: retail market dynamics and other key determinants. APA National Planning Conference, 1999.

·         Chang-Ing Hsu and Shwu-Ping Guo. Residential location choice in a n urban area with surface streets and rail transit lines. TRB Annual Meeting, 1999.