Data Collection
1 Data
Sources
2 Data Manipulation Some of the data layers obtained from S-Drive showed areas that were not needed for my project. The layers landuse.rst and ct96.rst had to be clipped to show mainly the Burnaby region. I used the WINDOW module in Idrisi32 to perform this clipping operation. The new clipped images I called burnaby_landuse.rst and census_tract96.rst respectively. I would later use these files as base layers for some of my analysis work. I found Idrisi32 Database Workshop to be an invaluable tool in expanding the tabular dataset I acquired. I added a new field called "pop_density" because this field was not provided with the original data. This field uses population totals for each census areas divided the census tract area, and so I used "Calculate Field Values" to do this. Having this population density field in my data table is invaluable since I could create a map of the population density distribution for Burnaby. The pop_density field also allows me to create more specific data layers relating to population density distribution later on. 
A step I took before using burnbus.vct was to transform the reference system of burnbus.vct from plane to utm-10n. This was because all of the other data layers I used were referenced using utm-10n and only burnbus.vct was referenced under plane. I ran PROJECT specifying that the output file uses the utm-10n reference file. This procedure successfully projected burnbus.vct into utm-10n so that it can now be overlayed on top of the other data layers. Before some spatial operations could be performed I had to rasterize the vector layers. This included rasterizing all the vector layers with the exception of burn_census.vct. I decided to use census_tract.rst instead because there were visual "holes" in burn_census after I rasterized it. Basically my digitizing of the polygon boundaries cannot be precise enough so that there were no gaps between polygons or that the polygons didn't overlap. The burn_census.vct however was put to use in allowing map displays to be generated when querying the datatable. 3 Digitized Data Layers burn_census.vct - This is a polygon vector layer that I digitized in Idrisi32 using its digitizing tool. This is a key base layer for all of the database querying that I performed. I decided to digitize this layer instead of relying mainly on the ct96.vct for several reasons. First of all, if I created my own polygon vector layer for Burnaby then I could link it to the database of demographic figures for the Burnaby census tracts. I used burn_census.vlx (a vector link file) to perform this link of the vector layer and the demographics datatable called data. Secondly ct96.vct included the whole Greater Vancouver region, more than was required for this project. It was not possible to extract only the Burnaby census tracts from the rest of the census tracts, therefore I needed a new data layer which only included Burnaby census tracts. This polygon layer is spatially referenced to the coordinates of burnaby_landuse layer so that it can overlay on top of the other. restaurants.vct. - I also digitized this point layer in Idrisi32. This layer provides the location of all chosen competitor fast food restaurants in Burnaby. I used allroads.vct as the guide layer to help locate the positions of the restaurants. I decided not to create a linked datatable to this layer since it might not really be worthwhile for the analysis I performed and the data that I could acquire. More on this discussed in the Problem Issues section. This point layer is also spatially referenced to burnaby_landuse.rst as well. |
