Geog 355 Final Project -- Methodology

All the analysis in this project was made possible by the use of IDRISI while the manipulation of data was done by ArcMap. For the purpose of analysis, raster format was used. To keep things simple, I will not go into detail of how to convert vector to raster. Below are some descriptions of how the spatial analysis was done in IDRISI.

Proximity to public transit

The most important factors in this project are the proximity to different transit services. For convenience I called that “Transit Factors” here. The first step in the project is to determine the distances from bus routes, bus exchanges, B-line bus routes, peak hour bus routes, and skytrain stations. Fig 1 shows how I determine the distance from these locations using the “distance” module in the macro modeler in IDRISI. Reclass is sometimes needed before calculating the distance because I want all the public transit maps to be in a Boolean format for easy manipulation.

MCE Analysis

The resulting maps from the distance determination process generate the continuous surfaces showing the distances from all the transit factors. We can now perform the MCE (Multi-Criteria Analysis). All of these factors were inputted into the Decision Wizard and the constrain for this analysis is the City of Vancouver raster surface I created. Since this analysis is based on distance and generally the attractiveness of public transit decreases with distance, I choose the “monotonically decreasing linear” function to describe these functions in our analysis. A series of fuzzy images of the factors I inputted as a result. After that, I calculate the weight of each transit factors by AHP (Analytical Hierarchy Process). The two images on the right (Fig 2 and 3) show how I perform the weighting process and the weight of each factor.

The resulting map from the AHP shows how likely people will use public transit based on the distance and the attractiveness of the type of transit services.

The second purpose of this project is to determine which areas need improved public transit services. This analysis was done in Macro Modeler and Fig 4 shows the macro modeler. Notice that the modules “group” and “areas” are used here to determine the areas that show a low score (low ridership of public transit) in the previous MCE analysis. Areas with low score would be identified as a need of improving transit. Areas would be calculated from the groups of low score. The largest area groups would be the communities that need better transit services.

The last step of the analysis is to determine areas that are suitable for higher density development. The factors for this analysis are to have a community of high transit ridership, with low density population, and with low density residential development presently. This analysis was done in Macro Modeler and Fig 5 shows the process. The resulting image shows areas suitable to convert to higher density development.