Detailed Methodology - Creating a Vector Based Point System at the Parcel Level

This section lays out the methods that were employed in Arc Info to create the point system. It is intended for people who are familiar with GIS, and describes in detail every step taken to create a parcel level point system. The AML that was written to automate many of these tasks is available on this site.

Data Preparation

In order to compile large amounts of information into a single point value for each individual parcel it is necessary to combine all of the resource coverages into one large file. This was accomplished by unioning the coverages using Arc Info. Before this was possible, an attribute had to be added to each coverage with a unique name, and an integer of 1 or 0 depending on whether or not the resource is present in each particular polygon.

The polygon attribute table was then cleaned to get rid of any unnecessary attributes. The resulting file contained 40,000 polygons, and had information in it from 15 existing coverages. This task was competed in Arc Info because it is necessary to have the area of each polygon. This can be accomplished using an Avenue script in ArcView, but ArcInfo was employed in this case. ArcInfo coverages provide more flexibility for analysis and were necessary to run the AML used in this analysis.

This coverage was then unioned with the parcel coverage. The result was a coverage that had all the parcel polygons, which could be individually selected, and included all of the polygons for the different resource layers.

Finally, a copy of the original parcel coverage was made, which had the same unique identifiers as the unioned parcel coverage.

Automating the Weighting

An AML was written to automate the coding of different weighted values for each resource. This was accomplished by selecting all polygons where a resource exists in the unioned coverage. The reselect was possible because of the binary coding of the existence of the resource mentioned above. A variable was then entered into the program, after the user was prompted to enter the weight for the resource, and this variable was then calculated into all selected polygons. This task was then repeated for each resource layer.

Automating this portion of the process is critical to analyzing many different weighting systems, which are the variables in question in this analysis.

Totaling up the Points, and Normalizing for Area

The total number of points in each polygon was then determined using the calc command in ArcPlot, and this value was coded into an attribute for the total number of points.

Next, the percentage of the parcel that the polygon occupied was determined. This was simply by dividing the area of the polygon, by the area of the parcel, which was given a unique attribute before they were unioned together. The percentage of the area was then multiplied by the number of points in the polygon to calculate a parcel-normalized point value.

Coding the Parcels

The parcels were then coded by taking advantage of the unique identifier in both the unioned parcel coverage and the original parcel coverage. A loop was set up to run through each parcel. The polygons within the parcel were first selected in the union coverage, and the normalized point values were summed and stored as a variable. The same parcel was selected in the original parcel coverage, and the sum of the points was calculated into a field in the original parcel's attribute table. This solved the problem of the one to many relationship that exists when multiple coverages are unioned together.

Adding in the Growth and Size Points

Some of the resources that were included in the scoring system were alreaded encoded into the parcels before this analysis. This includes the number of houses at buildout that was used to allocate points for growth management and the results of the online survey of cultural resources. Adding these resources to the scoring system was accomplished by joining the coverages tables from these coverages with the parcel coverage used in the scoring system and adding in the point values given for each resource in parcels where the resource exists.

Taking Cost into Account

The next step involved calculating the area of the parcel in acres, and using this to determine the amount of money it costs per acre of land purchased. The amount of points in the parcel was divided by this value to give the final score: the number of points per acre per dollar spent.

The results of this analysis are available on this site both as static maps, and using Arc IMS.