Collecting the Data:
 

• Actual work on the project started with an exploratory look for the necessary data.  Fortunately being a year after the election there were several decent election data sets available.  This provided me with a platform to begin the project.  After reviewing the data in ArcView I was able to determine what other types of data might be useful to make this project interesting.

• Finding socio-economic data was high on my list because it is useful discerning trends or voting patterns.  It was actually quite difficult to locate socio-economic data at the county level in a format that could be converted for use in Idrisi without taking too much time.  In the end socio-economic, voting results,  voter turnout and rejected ballot data was successfully located and converted.

Image Conversion (raster and vector):

• All the images used for this project came in Esri Shapefile format, and needed to be converted into Idrisi raster images.  This was accomplished by turning on the spatial analysis extension in ArcView and converting the Shapefiles into grids.  After converting the shapefiles into grids, they were exported using export data source in the file menu, in ASCII raster format.

• Once the Shapefiles were converted to grid files and exported in ASCII raster format, they needed to be imported into Idrisi and converted into Idrisi raster images.  This was done using Import -> Software-Specific Formats -> ESRI Formats -> ARCRASTER, which is found in the file menu.  After converting the grid file into an Idrisi raster a vector copy needs to be made using Raster/Vector Conversion -> POLYVEC in the reformat menu.  A vector copy needs to be made so that the database can establish a vector link, which allows the data to mapped.

Databases:

• Data used to create the census and election results databases for this project came from several sources and had to be manipulated into a workable format.  Manipulation mainly revolved around editing the data so it that was consistent with how the raster images in Idrisi were arranged.  This meant figuring out how Idrisi was IDing the polygons (that had been rasterized in ArcView) and arranging the data accordingly.

• Arranging the data and merging data sheets was accomplished in spreadsheets using Microsoft Excel.  Data was manipulated / edited until all useful collected raw data was sucessfully included.  I did not completely merge all information.  I chose in the end to maintain two datasets for each level of data detail (census and results).  When the datasets were completed a copy was saved in Dbase format (.dbf).

• The Dbase versions were imported into the database workshop in Idrisi using Import xBase (dbf) in the file menu.  In order to use the data for spatial analysis the database had to be linked to a vector image (vector image corresponding to the correct level of data - county or state).  This was done with Collection Editor in the file menu.  A vector link file was created for each database to allow the databases to establish display links, so the data could be mapped.