Final Project Report: GEOB270 Report

For our final project, my team and I wanted to assess the relationship between income and crime per capita within the census tracts of Vancouver. Our goal was to construct two separate maps for income and crime per capita respectively to draw comparisons, to construct a heatmap to determine the highest crime area(s) in Vancouver, and to construct a geographically weighted regression (GWR) map to assess if there is a correlation between income and crime per capita.

In terms of project management, my team and I worked on the GIS analysis portion of the project together in order for us to be familiar with the analytical steps taken. Once we finished doing the necessary analyses, each of us were responsible for making a map easier to understand (via cartographic practices). We then discussed our separate maps in the report, and divided the other written sections of the report approximately equally.

One interesting thing we learned as a result of conducting the project was how to make appealing hotspot and GWR maps that conveyed information sufficiently using ESRI’s ArcGIS Desktop 10.6.1. GIS software. Also, we discovered how difficult it is to obtain publicly available data of the same feature (such as crime) with the same parameters from different cities. Initially, we wanted to attempt to map crime density for all of Greater Vancouver, but found great difficultly in locating comparable datafiles of crime points from the different municipalities, resulting in us deciding to focus more on studying Vancouver. Furthermore, we learned that when attempting to run tabular or spatial join, the CTUIDs (or any other data element) used to join two datasets need to be exactly the same, including the number of decimal points the CTUIDs have. During our analysis, one of our CTs was not joining properly because one of the CTUIDs ended with “.0”  while the other having no decimal points, resulting in ArcMap not recognizing them as the same CT. Therefore, we needed to make the appropriate change to our datasets manually to alleviate this minor but significant issue. Lastly, Google Drive was a very helpful tool for data management and sharing.

Using a number of manually sourced datum, is able to generate adjacent maps comparing income and crime per capita, hotspot maps of crime density, and geographically weighted regression maps of crime per capita as a function of income, resulting in a seventeen page professional report outlining the project objectives, methodology, results, errors, uncertainties, and next steps in detail.

Lab 5: Environmental Impact Assessment (EIA)

Environmental Impact Assessment Map of the Proposed Brohm Ridge Mountain Resort: EIA

Environmental Impact Assessment Map of the Proposed Brohm Ridge Mountain Resort (3D Hillshade): EIA3D

Environmental Impact Assessment of the Proposed Brohm Ridge Mountain Resort Project Memo

            British Columbia is world-renowned for having some of the most attractive ski resorts on the planet, with tourists flying in every year to ski or snowboard on its majestic slopes, and to enjoy the beautiful scenery it has to offer. With the 2010 Winter Olympic Games taking place in Vancouver, ski resorts have become an increasingly profitable investment. As a result, the Aquilini Investment Group and Northland Properties Corporation have sought permits for a proposed ski resort on the slopes of Brohm Ridge near Squamish and the Sea-to-Sky Highway. The Environmental Impact Assessment (EIA) of this $3.5-billion project has already taken place.[1] This memo will summarize my analysis and results, will outline my sources of errors, and will discuss further recommendations to explore.

            The analysis began with the acquisition of datum, much already available to me (complements of UBC Geography) and some (specifically Ungulate Winter Range areas and Old Growth Management areas) downloaded through DataBC. Using ESRI's ArcGIS Desktop 10.6.1, all the layers were converted to the NAD_1983_UTM_Zone_10N coordinate system once they were imported, after which the data from the Digital Elevation Model (DEM) was manipulated to differentiate the area that is at or under 600m elevation and the area that is above 600m. The Ungulate Winter Range areas, the Old Growth Management areas, and the Fish habitat riparian management zones were clipped to the project boundary, with the stream being given a buffer relative to its elevation (A narrower 50m buffer was provided in elevations higher than 600m, and a wider 100m buffer was provided at elevations below 600m). Moreover, a Union of the 3 mentioned layers’ area relative to the project boundary area was used to compare the sum of each layers’ area relative to the project boundary area. Then, the EIA map was constructed, using general landuse data alongside the 3 protection layers as well as the park boundaries, river and road. Similarly, a 3D hillshade EIA map was created with the snowline contour alongside the 3 layers as well as the river, road, and 20m elevation contours. Inset maps were added to both maps for reference.

            There are a number of important EIA findings to take into consideration with this project. 31.78% of the proposed area is in elevation equal to or lower than 600m which may be unsuitable for ski slopes as there may not be enough snow during the ski season. Individually, 7.89% of the project area has combined Mule Deer and Mountain Goat winter habitat, 6.78% of the project area has old growth forest, and 14.98% of the project area falls within fish bearing streams or within fish habitat/riparian areas around streams. Collectively however, 26.92% of the project area is protected in some way from the proposed project, meaning that 2.73% of the project area has more than one type of protection as certain protected extents occupy the same protected area. It is important to assess the impact of climate change on the snow line as well as on the environment as a whole moving forward in order to ensure an ideal ski resort location, to the benefit of investors and the environment alike. It may even be worth elevating the snowline contour as a provision for warming temperatures moving forward (assuming this will be the case). In order to undertake a more robust EIA, further analysis on rodent, avian, insect, and flora life should be assessed, as well as waste management infrastructure plans to name just a few variables worth considering.

            Sources of error in my EIA include a lack of complete and accurate data. The data may not have been up-to-date, resulting in a map that may be temporally inaccurate. Also, seasonal changes in snow cover and extent were not taken into consideration, especially with the added risk of climate change. Furthermore, errors in Topological Analysis, including Slivers, Overshoots, and Dangles may have been present. Moreover, recent changes in the proposed project may not have been taken into consideration in the analysis. Errors in the findings may be a result of a lack of considered variables during the analysis. Finally, human error is always a potential factor in any human-led analysis, as well as bias and assumptions that may have been introduced in the analysis, which would have cascading effects on the rest of the analysis and results.

            Project proponents are recommended to assess other areas in the vicinity where a ski resort can be built as Brohm Ridge Mountain is relatively saturated with dispersed protection zones that makes planning for a project of this magnitude somewhat difficult. That being said, some of the best available land for a resort appear to include a slope in the center of Brohm Ridge, encircled by the river buffer, or certain parts of the southern boundary as they have road access and are relatively clear.

            Moreover, Indigenous peoples of Brohm Ridge should be consulted, and their way of life should be considered. A more extensive EIA must consider any Indigenous traditional knowledge alongside its scientific findings. It is equally important to consider the degree in which the proposed project’s distribution of environmental benefits and detriments is equitable. Nonetheless, a strong precautionary principle is advised when applying the EIA into practice, one that limits discretion with required obligations as a response to red flags, and one that requires that potential long-term effects on the environment are prepared for.

            All in all, the proposed Brohm Ridge Mountain Resort project is one that must take into consideration a number of protected areas, and that will require further EIA in order to minimize uncertainty regarding its consequences to the environment. Two EIA maps were constructed, one with general landuse data, and one with 3D hillshade data, with both displaying the many protected areas. While it is clear that more than one quarter of the area is protected, a number of potential errors as well as limited variables considered signal for a more in-depth EIA to be conducted, with more variables to be taken into consideration.

Yassen A.
The University of British Columbia

[1] Please see dailyhive.com/vancouver/garibaldi-at-squamish-ski-resort-february-2019-project-update for more information.

Using a number of landuse types and protected zone boundaries, is able to generate a topological environmental impact assessment map of proposed ski resort projects using ESRI’s ArcMap Desktop 10.6.1 to assess the environmental viability of constructing ski resorts in a given area, resulting in the ability to support environmental impact assessments using GIS.

Lab 4: Housing Affordability
Working with Census Data

Testing different quantitative data classification methods for dwelling cost in greater Vancouver: lab4dataclass

As can be observed in the aforementioned map, different data classification methods greatly influence the map's results, thus influencing our interpretations of the data on the map. For instance, with housing cost maps using the Natural Breaks classification method, houses in greater Vancouver generally seem more expensive than maps using the Manual Breaks classification method. Similarly, maps using the Manual Breaks classification method seem to show a general trend toward cheaper houses than maps using the Natural Breaks classification method.

Likewise, maps using the Equal Internal classification method appear to show that there are an equal number of regions for any given budget, and that housing costs are distributed equally. Also, maps using the standard deviation classification method suggest that there are a proportional number of cheap (below average) and expensive (above average) houses, regardless if the distribution is in reality skewed to one side.

Data Uncertainty:

In the abovementioned maps, certain census tracts were not coloured as a result of containing no data. One reason for this is due to Statistics Canada’s rules on data suppression:

• No characteristic or tabular data can be released for standard areas with a population size under 40 people, and for nonstandard areas with a population of under 100.
• Clients cannot be provided a list of postal codes except postal codes with specified names in their request.
• Quantitative data will be suppressed if below a certain specified threshold (such as income) or if the number of data points is less than 4. Statistics will also be suppressed if an outlier exists that surpasses a given threshold.

In this regard, the way in which census tracts (CTs) were delineated is important to take into consideration. Statistics Canada tries to ensure that they follow permanent and easily recognizable features, that the population of a CT ranges between 2500 to 8000 people, that the population within the CT is homogenous as possible in respect to socioeconomic status, that the CT’s shape is as compact as possible, and that the CT boundaries respect other boundaries set by the government.

As for the CTs shown in the greater Vancouver maps, 19 data polygons have “0” recorded for shelter or median value of dwelling, likely because the data is suppressed by Statistics Canada since Indian Reservations cannot be seized legally by the Crown nor is it the property of a band member. These CTs make up about 4% of Metro Vancouver CTs (since there are a total of 468 CTs).

Some Pros of using CTs are that there are many census years worth of data available, that CT splits can still be compared to original CTs for historical comparisons as the user can still re-aggregate the splits, and that it is a relatively stable geographic unit. Some Cons of using CTs include that CT boundaries do not always respect census subdivision (municipality) boundaries.

Maps comparing housing costs in Greater Vancouver and Montréal: housecostVM

This data represents the different housing costs (in $CAD) per CT in Greater Vancouver and Montréal. The data was obtained from Statistics Canada (2016). ‘Shelter cost’ refers to median cost of dwelling for homeowner in 2015.

One significant problem with using housing costs for making comparisons of different places is that the comparison is almost meaningless as the maps do not inform their audience about living affordability between these different places given that different places have different living costs and pay workers differently based on these living costs, variables that are not included in housing cost maps. Essentially, purchasing power varies greatly based on region, even when comparing between places in the same country that use the same currency. Therefore, income (among other variables) should be incorporated to better assess living and housing affordability.

Map comparing housing affordability in Vancouver and Montreal: affordability

Affordability measures the extent to which the lower 99% can afford a house based on average income per household. It is a better indicator of housing affordability than housing cost alone because housing cost does not take the purchasing power of each place into consideration, while housing affordability does (as mentioned previously).

As seen in the abovementioned map, the housing affordability rating categories are “Affordable” (based on a Median Multiple of 3.0 and under), “Moderately Unaffordable” (based on a Median Multiple of 3.1 to 4.0), “Seriously Unaffordable” (based on a Median Multiple of 4.1 to 5.0), and “Severely Unaffordable” (based on a Median Multiple of 5.1 and over). The Median Multiple is a consistent indicator for measuring housing affordability, calculated by dividing dwelling cost by household income. These categories are determined by Demographia, a group owned by American urban planner and policy analyst Wendell Cox. Given that Cox is known to be politically active and prefers reinforcing existing infrastructure rather than increasing city densification, his audience should be cautious when observing the data and maps that Demographia puts out given that the use of different data classification methods can lead the audience to different conclusions that may serve such political and even economic agendas.

As for whether or not affordability is a good indicator of a city’s ‘livability,’ it depends on what is meant by ‘livability.’ If livability refers to purely economic variables such as purchasing power, revenue, and costs, then affordability may in a fact be a good indicator. However, if ‘livability’ refers to a broader understand of one’s ability to live within a city, then affordability is but one of many variables that should be considered, including access to food, disease levels, air pollution levels, radiation levels, temperature, the number of opportunities available, infrastructure and available amenities, the level of happiness, as well as important variables for those with special needs such as the level of accessibility for the disabled to name a few.

Using a number of data classification methods, is able to generate dwelling cost and affordability comparison maps between two cities using ESRI’s ArcMap Desktop 10.6.1 to best display living costs, resulting in the ability to support public policy endeavors regarding city livability with GIS.

Lab 3: Planning for flooding/Storm Surge
Spatial Analysis, Map Layout, Editing

Below are maps of the Greater Vancouver Area using 25-meter squared resolution Digital Elevation Model (DEM) data. Map A represents the varying elevation levels in Greater Vancouver by area. About 187.3 Km^2 worth of total area is at or below an elevation of 6 meters in Greater Vancouver. Map B represents potential flooding areas in Greater Vancouver based on a projected storm surge of 6 meters within 1 km of shoreline given the varying elevation mapped in Map A. Map C outlines the potential flood-affected highways and expressways (about 63.6 Km and 105.9 Km total distance of road affected, respectively) in Greater Vancouver with the previously mentioned specifications, which may be used as a flood route guide to assist first responders who rely on these roads. Map D, using 1-meter squared resolution data acquired with LiDAR, looks at the planned construction location of the St. Paul’s Hospital in the False Creek area, a location which also happens to be in a low elevation area (2 – 3 m). Sources of data error include the age of the map data used (may not be fully up-to-date), the map data’s (relatively coarse) resolution, loss of information as data was transformed from raster to vector, errors in classification and generalization, inaccuracy when drawing out the polygon representing St. Paul’s Hospital, and, of course, other human errors.

Map A: lab3_lowelevleg

Map B: lab3_flooding

Map C: lab3_roadsflood

Map D: lab3_falsecreek

Able to generate elevation and flood-risk maps using ESRI’s ArcMap Desktop 10.6.1 to assess vulnerable infrastructure such as disaster response roads and hospitals, resulting in the ability to support emergency response planning with GIS.