January 3, 2018 – Introduction to the Course

This course focuses on different applications of GIS, specifically, for the management of landscape ecology, health geography and crime analysis. These areas of study are linked by the five themes: patterns, processes, places, people and perspectives. The idea is based on Tobler’s First Law: everything is related to everything else, but near things are more related than distant things. Therefore, no aspect can be looked at individually, without taking into account the entire system in which in belongs to.

Looking more in depth at each of these terms, landscape ecology is the study of interrelationships between organisms and the abiotic patterns and processes. Here, studying space can determine the biotic and abiotic factors that create the observed patterns of distribution. Health geography is comprised of 3 main elements: disease ecology, health care delivery, and environment and health. The study of spatial patterns and perspectives help to better healthcare services. In crime analysis, patterns and correlations are determined to enhance operation and maximize resources to suppress criminal activity. By thinking spatially, links can be found in each of these three disciplines that explain patterns and processes that could otherwise be overlooked.

 

January 10, 2018 – Why is Geography Important?

“Why is geography important?” This is an important question to address since geography is an important element in different aspects of life ranging from science, society, economy, politics and more. This week, is aimed at the basics, looking at concepts and terms that arise from studying and using geographic data. Some of these include:

• the modifiable areal unit problem (MAUP)
• the scale, grain and extent of a study
• the nature of boundaries of a study area
• spatial dependence and heterogeneity

A fundamental problem with interpreting patters in data is scale. As the scale at which a problem is being tackled on changes, so too do the patterns that are witnessed. When looking at ecological phenomenon, for example, the characteristic scales, which are animals’ typical habitat or range, can be used to determine an appropriate scale. Likewise, for crime analysis, the characteristic scales can be identified depending on the nature of the crimes committed. For instance, offenders who commit crimes near their homes can be mapped using a neighbourhood scale.

“Applied challenges require the interfacing of phenomena that occur on very different scales of space, time, and ecological organization. There is no single natural scale at which ecological phenomena should be studied; systems generally show characteristic variability on a range of spatial, temporal, and organizational scales.” – Levin, 1992

So, why does scale matter? This is because context is key to understanding geospatial patterns. Changing a spatial extent can shift the magnitude of correlations. For instance, when areas are clumped together, it is more likely that some finer resolution data will be lost, or reinterpreted as something else in the courser analysis, therefore, an area that might be a hot spot on a larger scale, might be generalized to a more neutral zone on a smaller scale map. This is especially true when looking for spatial autocorrelation.

Spatial autocorrelation refers to the phenomenon where the mere presence of some quantity of a variable in a sampled unit makes its presence in neighbouring sampling units more or less likely. In cases where similar values appear together, it is known as a positive spatial autocorrelation. For negative spatial autocorrelation, dissimilar values appear in close association or similar values are maximumly dispersed.

Spatial autocorrelation goes hand-in-hand with kriging. This is a geostatistical method developed to model natural random irregularities in sampled values. It is composed of three components: structural (constant trend), spatial correlation, and random noise. Taking all three, it produces an output that expresses the confidence of the spatially autocorrelated surface being analyzed.

This brings us to the modifiable areal unit problem (MAUP), which consists of two parts:

1. the scale and aggregation problem: where the same set of data can give two different results depending on the spatial resolution. Also, there are various ways in which spatial units can be grouped at a given scale, skewing statistical results. Simpson’s Paradox shows us that sometimes other elements impact a study using aggregated data. If there is varying correlation between different variables, then it may be hard to obtain reliable estimate of true correlation.
2. the specificity of the results to the data units used: this means that the results that are produced cannot be generalized to make extended conclusions. The results are only true for one particular case at one particular scale. Changing the spatial resolution or the location could change the results. For instance, if a study found that on average people over the age of 35 living in a certain neighbourhood have on average three children, it does not mean that on an individual scale, each person in that neighbourhood has three children. Certain families can have more or less and in a larger scale, this averages out.

So, through all the problems mentioned, it is important to take into account scale, space and place when assessing and analyzing a problem.

 

January 17, 2018 – Understanding Landscape Metrics: Patterns and Processes

Landscape ecology deals with the reciprocal interactions between spatial pattern and ecological processes that occur on landscapes. Ultimately, it is an understanding of processes through the examination of form, which might not always be simple to identify given that there can be feedback between the two. The fundamental assumption is that things do vary among locations and in relativity to other things which lead to important consequences.

It is easy to forget that observed patterns in cartographic depictions are only one of the possible outcomes generated. This is where statistical analysis comes in (we will touch more upon this in next week’s lecture). It looks at the probability of a pattern being generated by a particular process.

There are several methodological considerations that must be taken into account when looking at landscape ecology. The first relates to spatial autocorrelation and is referred to as first-order and second-order processes. First-order processes refer to patterns that develops due to a response to an environmental factor. Second-order processes results from the interactions between objects and events themselves. Then there is stationarity which considers a process that determines the placement of an object does not change over time.

The processes are abiotic, biotic anthropogenic and disturbances. The interactions between them are what govern landscape form. Some of the abiotic factors include climate, topography and soils. Disturbances can be anything from past and present human settlements to natural disturbances and succession. These are complex systems with multiple processes making them hard to analyze, especially when you throw in varying temporal and spatial scales.

 

January 24, 2018 – Statistics

This lecture provides a review on statistics as it is important for understanding models and the accuracy of results. Some statistical terms were skimmed as refreshers, including: measures of central tendency, measures of dispersion, sampling frames, and data measurements scales (nominal, ordinal, interval, ratio). The focus, however, was on regression which is the relationship between one or more explanatory variables, and one or more responsive variables. This relationship can be shown in a linear or multiple regression analysis to interpret results of grouping analyses and correlation matrices. One specific type of regression is the Geographically Weighted Regression (GWR) which uses this statistical tool in a spatial context. GWR is determined by examining neighbourhoods of points using distance decay functions. The determined coefficients are allowed to vary spatially, and a weighted matrix is defined for each point to model the relationships. This is a valuable tool for studying in a wide array of human and environmental processes that interact over space and time.

 

February 7, 2018 – What is Health Geography?

“Geography and health are intrinsically linked. Where we are born, live, study and work directly influences our health experiences: the air we breathe, the food we eat, the viruses we are exposed to and the health services we can access. The social, built and natural environments affect our health and well-being in ways that are directly relevant to health policy.” – Drummer, 2008

Given what Drummer (2008) states, our health is a manifestation of our geography. The basis for health geography, was medical geography, which is the application of geographical perspectives and methods to the study of health, disease and care. By looking at the similarities and differences of both fields, an amalgamation can be made for tackling problems in our current society. The current focus of health geography is access, looking at social, economic and political influences.

A key term here is environmental justice. This is the equal and fair treatment of all people regardless of ethnicity or national origin. This is, however, not always the case. Large populations of immigrants and other marginalized populations are often forced to live in areas that experience higher pollution. This is may be due to a number of different reasons, such as economic constraints and social factors, but, inevitably, this puts these people at a higher risk of developing health problems. This dynamic relationship shows that a person’s health and risk of contracting illnesses are inherently linked to power relations in society.

 

February 14, 2018 – GIS in Health Geography

As we continue to explore health geography, by looking at some of the GIS applications in this field. Here we look at:

• spatial epidemiology
• environmental hazards
• modelling health services
• identifying health inequities

The bulk of the focus was on epidemiology which is concerned with understanding spatial variation in disease risk. Here we looked at factors such as population distribution and movement, scale of study, and individual characteristics as a framework for analysis. Some of the issues that come with this type of analysis include spatial misalignment and sources of uncertainty.

Environmental hazards look at exposure to certain risks and the patterns of clinical outcomes. For instance, being downwind of a chemical plant leads to exposure to air pollution and detrimental effects to a person’s health. These health events can then be mapped and analyzed for spatial patterns.

This leads us to modelling health services. It is important that medical services can be accessed even in remote places. In Australia, this is modelled through an index of accessibility for all populated places that are outside of metropolitan areas. The importance of this is emphasized in a study done on infant mortality rates. It showed that the highest ratio of infant mortality occurs in remote areas while the lowest are in metropolitan areas. There is a further inequity when comparing Aboriginal children to non-Aboriginal children, with the difference being almost 3 times greater for the former. The inequity discussed in the previous example is just one of the limitations of healthcare. Some of the other socio-economic factors that result in inequities include: gender, education and unequal pay.

Through interdisciplinary factors, disease predictions have become linked to geography. Therefore, through the use of GIS, health and population data can be georeferenced to better incorporate environmental and anthropogenic factors.

 

March 14, 2018 – Is Crime Related to Geography?

As this course tapers to an end, we look at the final topic of crime as it relates to geography. Crime analysis is the qualitative and quantitative study of crime and law enforcement information in combination with socio-demographic and spatial factors to reduce and prevent crime, and create organizational proceedures. GIS aids in crime analysis in a number of different ways, the foremost being to visualize crime occurrences.

Three theories that are used in crime analysis include:

• the Routine Activity Theory: this is based on the activity of individuals, looking at the spatially non-random patterns of socio-demographic and economic characteristics of people. Hence, criminal activity is not random. Therefore, we can predict where crime would most likely occur.
• the Rational Choice Theory: this rationalizes a criminal’s thought process as they commit crimes, using the information to gather evidence and convict a felon.
• the Criminal Pattern Theory: this is based on the observed patterns of criminal activities which give insight to predicting future offences. This geographic profiling enables the analysts to look at the most probable locations of where an offender resides.

Several applications for crime analysis include tactical, strategical and administrative crime analysis. In a nutshell, these different applications work towards present and future crime prevention and are put towards making policies that are beneficial for the community.

 

March 21, 2018 – The Use of GIS by Fire Departments

Fire departments in municipalities constantly gather and do real time analysis on data in order to be better prepared for dealing with emergencies. GIS analysis can allow municipalities to make policy decisions, such as locations on new fire stations or fire hydrants. It can also aid in determining dispatch hot spots and ensuring that fire trucks are relocated to areas that experience a high volume of emergency calls.

This week’s lecture uses the Calgary Fire Department (CFD) as a case study for how GIS is incorporated to enhance emergency services. In Calgary, fire departments follow the National Fire Protection Association’s codes and standards, using GIS as a means to improve response times in the city and plan for future developments. For example, the CFD collects data on monthly calls, property loss, fire-related deaths and fire stations. The CFD also analyzes risk by collecting survey data, vulnerable structures are determined due to exposer due to proximity, and accessibility to emergency vehicles. All this information can help to prevent disasters by ensuring that enough fire trucks are rerouted to areas that are identified as hot spots, minimizing response times. In this way, GIS becomes a useful tool that affects everyone in the community.