Introduction
Everyone has a dependency on some sort of transportation in today’s society. Whether it is automobile, bus, subway, train, or aircraft, people always have places to go and things to do. An important factor to transportation is time. P eople ideally would like to take as little time as possible, the shortest route, if you will, to their destination. Another important factor is cost. Individuals would like the cheapest route, relative to how they value the cost variable. Nearly every business or organization must encounter these issues, whether implementing a trucking freight route strategy, designing a salesman’s district territory, air traffic control, or public transit; routes and networks are an important part of the infrastructur e of transportation.
Client Definition and Requirements
This paper will analyze the effectiveness of network analysis in a GIS environment, using Ottawa’s "Near West" area, comprised of the western fringe of the downtown area to Woodroffe Avenue. The client and their needs will be defined, a network
will be created and analyzed for the area. Vector and raster distance functions will also be applied and compared for effectiveness. The software used in this study is ESRI’s ArcView3.1, with the Spatial Analyst and Network Analyst extensions loaded.
The client is a person who lives at 563 Melbourne Avenue in the Westboro area of RMOC (See Figure 1). The client does not own an automobile, and dislikes multiple bus transfers to get to her destination. Most of her personal and profess ional commitments are
within Westboro; as a result she prefers walking along most streets and taking public transit along the Transitway. She would like a ‘best route’ analysis so she could more effectively plan her itinerary.
Figure 1 - Westboro, Ottawa's 'Near West'
Figure 1 shows the location of the client’s house, and the three stations where public transportation stops along the Transitway.
Source Data
The framework data for this study is a section of a street network file of the RMOC covering the area. An ESRI shape file is also supplied of the Ottawa River for point of reference and perspective.
Acquisition / Input
A great deal of time and effort was allotted to the preparation of the correct geocoding parameters, in order to create a strong network analysis model. After assessing the client’s needs, the street network file was opened in ArcView, in units meters. An address DBF file was created which contained the three intersections of the subway stops, intersecting at the following locations:
In order for geocoding to take place, the street network’s geocoding parameters first had to be declared and indexed through the Theme | Properties menu in preparation for geocoding. When the geocoding function was applied to the network theme and the address DBF file, the tolerance had to be set to low so as not to be overly stringent and return the matches desired. A point location file was created which corresponded to her the locations of the three Transitway stops outlined earlier, using ArcView’ s Geocoding function. The client’s residence was then added manually to the theme and referenced in
the linking table. At this point, we have created the parameters for the client’s needs in reference to her residence and stops along the Transitway.
The street network file was then edited to portray distance and time values as per the to and from routes of the table. A "Distance" field was created and populated with data returning the length of the shape of each record. The formula use d when editing the table in ArcView was:
Distance = Shape.ReturnLength
..which ran a built-in function and supplied distance information for each street in the network. A "Minutes" field was then created and calculated to return the time, in units minutes, taken to use the respective street. The values were we ighed and formulated to cater to the client’s desire to walk everywhere but the Transitway, in the following manner:
This outputted the adequate fields and records in the street network table to effectively apply a network analysis for the client Two fields; "F_Elev" and "T_Elev" were added to address the elevation parameters of the network. Norm al route or cost routes were assigned a default value of zero (0).
The newly added fields were then manipulated to block pedestrian access to the Queensway and allow access to the Transitway only at Transit Station specified earlier in the paper, setting up overpasses. These routes were edited to a value of one (1). Accessibility is then translated as per Figure 2:
Figure 2 – Overpass Concept Weighting
At this point, the data is formatted and ready for network analysis with the Network Analyst extension.
Application
A service area theme for the client was created, based on the client’s house location and the derived network. The service area was classed into intervals or 5, 10, 15, 20, 25 and 30 minutes. The result is illustrated in Figure 3. Wes tboro Station and Tunney’s Pasture
were omitted from the calculation so as generate a continuous network coverage to Lebreton.
Figure 3 - Service Area Network Analysis
The polygon vector network appears as rings of concentric zones relative to the client’s residence. Distance increases with darker polygons. The routes, in contrasted colour, also have the same effect due to the reclassification of the time intervals . Distance increases with lighter coloured lines. The pattern appears irregular due to the definition
of the network criteria, so that roads that seem farther than others are included in the service area. This is due to walking / public transit speed parameters, and the overpass boundaries. One anomaly that occurs is at an area close to the Queensway area (Figure 4). The client appears able to cross the Queensway by proceeding around the highway. The network table suggests that the highway is assigned another street name, which she can walk on.
Figure 4 - Queensway
The travel times derived from a network analysis differed considerably when compared to those derived from an as-the-crow-files algorithm. ArcView’s raster-based Spatial Analyst extension was used to create a distance model not taking into account rou te attributes. The Spatial Analyst Find Distance function was used, again based on the
client’s residence, and reclassed into the same time intervals. The map calculator was then used to represent the time intervals as needed.
Figure 5 - As-the-Crow-Files Distance Analysis
The map in Figure 5 can shows travel times derived from Spatial Analyst. It is evident that Spatial Analyst does not take into account cost functions as efficiently as Network Analyst does. Spatial Analyst’s zones of interest "fork" out in a circular motion based on the inputted point theme.
Analysis
As a result, in this case study, the client’s needs are met by creating a network schema with ArcView’s Network Analyst Extension. The Spatial Analyst function faired poorly in comparison in this environment, weighting most of it’s crit eria on distance factors alone, not taking into account variables such as boundaries or user-defined costs. For example, the client would end up walking on the Queensway, or out into the Ottawa
River for that matter. In this case study, the Network Analysis functions proved better suited to take into account such factors as speed of mode of transportation (walking vs. public transit), overpasses (highways), and desired length of time as inte rvals for the client.
Delivery
In conclusion, the client receives a customized network route map, as per her requirements. A floppy disk of the final data is attached to this report, with an ArcView Project file (tom_kralidis_a3.apr) for the client’s review.
ACRONYMS
ESRI – Environmental Systems Research Institute
GIS – Geographic Information Systems
RMOC – Regional Municipality of Ottawa-Carleton
DBF – dBase file format