A GIS for Public Transit

Abstract

The Corpus Christi, Texas, Regional Transit Authority and Texas A&M University-Corpus Christi have developed a geographic information system that includes the street maps for the three county service region, the route system, and the bus stop locations. These maps are used together with US Census block and block group information to perform communication, analysis, planning, and service assurance. A GPS survey is currently underway to support automatic vehicle location and data collection.

Background

The work of a regional transportation authority is inherently geographic. Routes are plotted on maps according to the potential to generate ridership. Ridership statistics needed for validation and maintenance of ongoing routes are anchored to the location of these routes and associated stops. The ability to work with maps, routes, and demographic information is essential to the management, development, and maintenance of a regional transportation system. While this need has always been recognized, until recently high prices and operational complexity kept the automation of these functions from being fully realized by all but the largest systems. Developments in geographic information systems (GIS) technologies now allow cost effective automation of many key service activities for RTA systems of varying sizes.

Transportation agencies are subject to federal government statutes plus the laws and regulations of state and local agencies, and/or their own board of directors. Much of the response to the requirements generated by these entities is in the form of maps and/or statistical data. Without the availability of a computerized geographic information system, maps must be drawn by hand, and the presentation of associated statistics is virtually meaningless to those unfamiliar with what the numbers represent.

The Corpus Christi, Texas, Regional Transportation Authority (RTA) was operating in just such a manual mode in 1994. Prior to the establishment of a GIS, maps were drawn by hand against a lighted backdrop. Statistical analysis was done manually using a primitive database, but mostly with brain power and an immense amount of time. Local RTA officials felt that the establishment of a GIS would be beneficial, but would require a larger budget and greater capabilities than available "in-house". Texas A&M University - Corpus Christi (TAMU-CC) was approached about the possibility of contracting for the development of a site specific GIS to meet their requirements.

Contracting by the RTA with TAMU-CC was felt by both entities to be far more economical than hiring custom designers and programmers. The RTA chose to use ARCVIEW in the development of their GIS because it is both a basic analytical tool and a powerful mapping system, as well as being fairly inexpensive. TAMU-CC used ARCINFO for the development of the base map and the establishment of the routing network. This paper will cover the steps involved in the development of the Geographic Information System for the Corpus Christi Regional Transportation Authority.

Creation of the Base Map

The resulting prototype immediately illustrated the potential of this GIS. The ease with which the RTA was able to create maps and the ability of these maps to convey information were of great satisfaction to RTA officials. Their only dissatisfaction was with the extremely slow draw time.

Improvements to the Base Map

The original base map required vast amounts of computer storage space. The map was composed of multiple arcs between intersections when only one arc was needed. Often, these multiple arcs would represent curves, and in the majority of curves were a multitude of vertices. In many cases, a single curving arc would have in excess of 100 vertices. (The arc highlighted in the "multiple arcs" example contained 132 vertices. The GENERALIZE command reduced the arc to 5 vertices.) Additionally, duplicate arcs would occassionally be found. Texas Department of Transportation maps gave emphasis to major thoroughfares by increasing their size relative to other streets. (In the example, the blue lines indicate where the arcs should have been.) This caused street portions that lay under expanded streets to be deleted. In many cases, the deleted street portions were needed within the route system. Attention had to be given to moving the expanded streets into a more representational position and adding to the map those streets or portions of streets that were missing.

Another vast editing job was begun to reduce the number of arcs in the map and to make necessary corrections to the street network. The resulting reduction in storage requirements for the excess arcs and vertices was found to significantly improve the drawing time on subsequent maps.

Attribute Transfer

Creation of Route Buffers

In order to use this information for ridership analysis, buffer zones around the routes first needed to be established. In general, the majority of riders will walk no more than 1/4 of a mile to catch a bus, and a 1/2 mile boundary will gather in most of the remaining ridership. A buffer zone defined around a route and laid over the blocks will include some partial blocks. Therefore, it is necessary to establish what percentage of block data should be apportioned to partial blocks. By using the fractional area and population data of the blocks included within route buffers, a higher quality proportional allocation of block group information can be made. The greatest benefit of the GIS to the RTA has been this ability to analyze their routes in correlation with census information.

An Arc Macro Language (AML) program was written that creates three buffers around each route. The buffers are at distances of 1/4, 1/2, or 3/4 of a mile on either side of a route. Each buffer is then used much like a cookie-cutter to "cut" that portion of the block data from the census block files. The AML then determines what percentage of the whole block area is within the buffer block. (In the example, the darker the color, the denser the population.) RTA officials are able to use that percentage to allocate the portion of census data to the blocks within the buffer and thus determine to what extent they are meeting service guidelines. Officials were quite satisfied to find that they surpassed service guidelines.

Geographic Positioning

Conclusion

Acknowledgments

Authors