 |
|
|
|
|
|
|
 |
|
|||||
 |
 |
 |
 |
 |
 |
|
Data Manipulation
The filtering of the dataset was the major task of this project, as so many types of sites were included in the dataset. This filtering process had to be completed manually as the typology section of all 2200 tuples was a single column with several attributes listed in this column. As there were also several other columns that used some of the same terminology as the typology section (i.e. directions to the site), it was impossible to filter the data using any automated programming. The filtering was done by alphabetization and selection groups for manual deletion. Once the undesired groups had been removed (i.e. historic buildings), the sites were selected for by description of habitation site or cultural depression. The term cultural depression can include cache pit features but it was known that in later steps the spatial extent of the site would be used to select of areas that were small compared to the described habitation sites.
Once this was done, the chosen sites had the irrelevant data removed, such as descriptive location, ownership, site recorder, founder etc. This database file was then joined to the shape file using a limit outputs to those with primary keys in the joined database. The output was the final shape file that was going to migrate to IDRISI Kilimanjaro. The formatting time of the Archaeology Branch data required to complete this project severally limited the extent of the analysis and contributed to a great deal of frustration.
Other data sets used include a clipped version of the Cascadia DEM and a clipped portion of the rivers layer from the BC 2 Million dataset, both from the datawarehouse directory on the SIS drive. These were manipulated to fit into the scheme of the cost surface analysis that was completed. These maps were projected using BC Albers and used the coordinate system NAD83. The ability to overlay these maps percisely was a problem using IDISI but the reluting layers do not deviate a great deal from the visual understanding of the landscape.
These maps were all converted to the resoltuion of the Cascadia DEM layer, as it had the coarsest resoltuion and avoided increasing the resolution of any of the data. This did cause some problems in the processing stage and will be elaborated on further in the problems section.