2Data Issues

This project uses data from several sources and then manipulates
this data in order to provide meaningful results. I tried to make sure
that all/most data was accurate to 1998.
 

Data Sources

i. The base map of the GVRD was available
on the Geography server. It was originally taken from
the Canada Census data-set.

ii. All information on Crime in the Lower
Mainland was taken from the SFU Department of
Criminology's Research Laboratory, and specifically
from the BC Office of the Attorney General's
piblication entitled: "BC Crime and Policing Statistics".
This information is accurate to 1998.

iii. The information on the GVRD's driving times
to Richmond, and Average House Prices comes from the
GVRD's website and is accurate to 1998.

iv. A big part of my project was the GVRD
Streets Network coverage. This originally came from the
Canada Census data-set and was later transformed to
properly overlay on top of the GVRD municipal-areas
base-map coverage.

v. The information about the location of
Residential (Single Family, Townhouse, and Low-rise
Apartment) living areas in the Lower Mainland came
from a physical map from the GVRD entitled
"GVRD Land Use" that was published by the GVRD
in 1996. I borrowed this map from the library and then
manually plotted the residential areas onto a new
coverage in ArcView, using manual (heads-up)
digitizing techniques.

vi. The information on transit routes came from the
Coast Mountain Buslink website. Using address and
cross-steet locations on my Street Networks coverage,
I created a new coverage entitled "Skystrain Stations"
which was a point-coverage locating Skytrain stations. I
also created a new point coverage entitled "B-Line stops"
which demarks the stops for the No.99 B-Line express
bus, from Brentwood Mall all the way to the UBC area.

vii. Information about Parks in the lower mainland
was provided by a GVRD road atlas (1998
edition). I, using street networks' intersections manually
digitized parks in relevent neighbourhoods. Note that
the Parks information in the original Canada Census
Street Networks File was not used since it only contained
the major parks and not the smaller ones.

viii. Information on Shopping Areas was taken
mainly from my Bus Routes map which lists the major
shopping malls in the Lower Mainland. I also wanted to
include the locations of major supermarkets, so I used
the 1999 Metro Vancouver White Pages to note down
the street addresses of all stores of the following four major
supermarket chains: Safeway, Superstore, Save On Foods,
and IGA Stores. This all went into a new points coverage
entitled "Shopping locations" where I manually digitized
(using street locations) the locations of all the
stores which happened to lie on my preferred residential
areas. From this points coverage of relevent shopping locations,
 I made 0.6km (short walking distance) polygon buffers around
shopping areas and thus created a new polygon theme
"Shopping area buffers". These areas denoted parts of the
Lower Mainland that were in close walking range of shopping
locations.
 

Data Manipulation

i. The base map of GVRD included several areas
which I did not want to include in my analysis since they
do not represent appropriate living areas for Leigh. These
deleted areas include: Lion's Bay, Bowen Island, etc.
 ii. The GVRD street networks coverage was
transformed to fit on top of the GVRD municipal areas'
base map coverage by Lorraine. I also removed any
streets that did not occur on my GVRD base map. I
then did a very complicated query on the Street Network
information to select only the streets from all the entities.
(The coverage also contained entities such as parks,
railway lines, water bodies, etc.) The exact query
was:

([Type] <> "") or (([Addr_fm_le] > 0) or ([Addr_to_le] > 0) or ([Addr_fm_rg] > 0) or ([Addr_to_rg]  > 0)) or ([Name] = "*WAY*") or ([Name] = "*RAMP*") or ([NAME] = "*PKWY*") or ([NAME] = "*MALL*") or ([NAME] = "*ALLEY*") or ([NAME] = "*FRWY*") or ([NAME] = "*DRIVE*") or ([NAME] = "*BYPASS*") or ([NAME] = "*VIADUCT*") or ([NAME] = "*DIVERSION*") or ([NAME] = "*VIADUCT*") or ([NAME] = "*ROAD*")

This removed about 8,000 entities from the
Street Networks coverage, leaving behind just the approx.
52,000 streets. The query basically selected all entities that
had either no name, no type, no address-range, or
no "street"/"road"/etc. title in their name. Then I
inverted the selection in my Street Networks table
and simply generated a new shape file which resulted
in just the roads network within it.

 The final manipulation to my street networks
coverage was the Geocoding of all the streets' information
into ArcView, so I could perform a search for an exact
street number and address or cross-section. This
was very easy to do as I just had to fill-in the appropriate
fields in the 'Geocoding' area of my Street Networks
theme's properties and then let ArcView do the Geocoding
on its own. (That took about 25 minutes of processing
due to the size of the streets network coverage)

iii. The transit information required quite a bit
of manipulation as well. I first made 1km. (walking-distance)
buffers around the Skytrain Stations and B-Line bus stops.
This went into a new polygon theme entitled "Transit
Buffers". I also manually entered (using street intersection
location from my Street Networks coverage) areas around
certain bus routes that would allow Leigh to quickly get
to her workplace. These bus routes were:
 

 Bus 401 and Bus 407 (These two buses
 stop directly in front of Leigh's workplace
 and thus are perfect for her!)

 + Bus 410, 403, 404, 420, 406 (These
 buses come within walking-distance to
 Leigh's workplace and so are worth considering.

 + Any buses directly crossing the path
    of either bus 401 or 407 (Since Leigh
 could then transfer onto the two direct
 buses and quickly be at her workplace)

iv. The Crime information had to first be manually
entered into the GVRD base map's table into new
fields. The crime information came in actual numbers
of particular crimes in all the GVRD municipalities
within the Crimes document. I added together all the
human crimes and all the property crimes in each
GVRD municipality manually and then entered these
into the two fields (HUMAN_CRM and PROP_CRM)
in the GVRD municipalities' base-map coverage. Once
there, this crime information was again summed together
and normalized to the population of each municipality,
thus creating the final crime table "CRIME_AMOUNT".
 
 

Note: For some data manipulation I needed to do
Polygon overlays between two polygon themes, and
I could not figure out how to accomplish this in
ArcView. (That is, do in ArcView what one can do using
the "INTERSECT" command in Arc/Info.) But I did
eventually realize how to do this. The solution is provided
by the "Geoprocessing" extension to ArcView which
clearly lets the user intersect two themes to create a
third polygon theme, via its "Geoprocessing Wizard"
menu-item. Thus... problem solved!