Tutorial 1: Heart Disease Rate – 2016 Southern U.S. Hot Spots by County

My single hot spot map (HeartDisease_Elizabeth_Lee) represents the 2016 heart disease rates (deaths divided by population) of Southern U.S. counties by showing hot spots (and cold spots and non-significant areas).  The CDC Wonder Heart Disease data was obtained from http://wonder.cdc.gov and brought into an ArcGIS file geodatabase.  I created two models: one to analyze the data into feature classes for each year and one to generate hot spot analyses for each year.  These hot spot maps were then animated.

In my 2016 map, cold spots can mainly be seen in Texas, Kentucky, Tennessee, Virginia, Maryland, North Carolina, South Carolina and Georgia.  Hot spots can mainly be seen in Oklahoma, Arkansas, Louisiana, Tennessee, Mississippi, Alabama, Florida and West Virginia.  In general, hot spots tend to cluster around the central Southern U.S. area.

Animations give viewers a sense of interactivity as images are seen as having movement.  By animating a heart disease hot spot map, the audience can see any changes, movements and patterns.  This may be particularly useful to professionals in the healthcare industry such as cardiologists or medical/health geographers.  They may be able to infer something about how the environment correlates with heart disease rates.  However, further research will have to be done in order to make conclusions regarding causes of heart disease and any patterns noticed.  Other potential explanations for deaths by heart disease besides environment may include genetics or lifestyle.

Lab 1: Edmonton, Alberta Land Use Change, 1966 to 1967

Executive Summary

Landscape changes around Edmonton, Alberta were analyzed from 1966 to 1976.  FRAGSTATS metrics were evaluated in order to generate a transition matrix to demonstrate land use changes over the time period.  Analyses show more urban built-up areas, less cropland, less non-productive woodland and more productive woodland from 1966 to 1976.  The city of Edmonton should consider a food and land policy taking into account the needs of its citizens, urbanization and future developments.

Introduction

When looking at land use, we must divide the different categories in order to strike a balance between quality of life, social welfare and other issues (Hansen 1984).  Urbanized lands comprise of residential, commercial, industrial and other areas (Hansen 1984).  As populations grow, people expand into formerly rural areas, deconstructing and reconstructing areas of land use (Hansen 1984).  Besides population, factors like age of first marriage, divorce rates, greater lifespan, wealth, and mobility affect households and thus urbanization (Hansen 1984).  The province of Alberta has “relatively large urban areas” (Hansen, 1984, p. 68).  But if every decade just 1% of agricultural land is lost to urbanization, all of the “good quality land” in Canada will be gone (Hansen, 1984, p. 79).  So, evaluating land use is significant in urban planning and ensuring a brighter future for the city of Edmonton.

Results and Discussion

From GeoGratis, Canada Land Use Monitoring Program (CLUMP) data were downloaded.  The vector files were converted to raster files with a 100 meter resolution.  After generating class descriptors files for each year, FRAGSTATS analyses were conducted.  Class and landscape metrics were evaluated.  A transition matrix was generated using a Microsoft Excel PivotTable to observe how land uses in 1966 changed over time.

Analyses show more urban built-up areas, less cropland, less non-productive woodland and more productive woodland from 1966 to 1976 (Fragstats_Map_Elizabeth_Lee).  99.96% of urban built-up areas in 1966 remain in 1976.  However, 21.06% of mines, quarries, sand, and gravel pits; 10.85% of non-productive woodland; and 14.12% of outdoor recreation areas were transformed into urban built-up areas.  Only 82.34% of cropland remains cropland from 1966 to 1976.

Only about 8.53% of non-productive woodland remains from 1966 to 1976.  Over half (51.62%) is transformed into productive woodland.  63.76% of productive woodland in 1966 remains in 1976.  But, 51.62% of non-productive woodland and 32.34% of unimproved pasture and rangeland become productive woodland.

Conclusion

The great increases in urban built-up areas a rising issue for Edmonton.  This is especially concerning due to decreases in cropland.  How will the city cope with the larger populations that likely correlate with urbanization?  As more and more people arrive in the city, there are higher demands for housing, services, and of course, food.  Many cities in North America see conflicts between residents’ demands for “sustainable urban food systems” and municipal development (Beckie, Hanson & Schrader, 2013, p. 16).  This is why the city of Edmonton should consider a food and land policy taking into account the needs of its citizens, urbanization and future developments.

Literature Cited

Beckie, M.A., Hanson, L.L. & Schrader, D. 2013. Farms or freeways? Citizen engagement and municipal governance in Edmonton’s food and agriculture strategy development. Journal of Agriculture, Food Systems, and Community Development 4(1), 15–31. http://dx.doi.org/10. 5304/jafscd.2013.041.004

“FRAGSTATS METRICS.” Accessed February 1, 2020. http://www.umass.edu/landeco/research/fragstats/documents/Metrics/Metrics%20TOC.htm

Hansen, J. A. G. 1984. Canadian small settlements and the uptake of agricultural land, 1966–1976. SOCIAL INDICATORS RESEARCH 15(1): 61-84.

Lab 2: A Child’s Language Skills in Vancouver

The relationship between a child’s language abilities and other variables to the child’s neighbourhood was examined (Elizabeth_Lee_GWR).  Exploratory regression was used to determine the most important variables to use in the geographically weighted regression (GWR) analysis.  With those determined variables, generalized linear regression (GLR) was performed to calculate the statistics correlated with the variables.  Then GWR was conducted to examine the spatial relationships with the variables.  Box-plots were then generated after using a spatially constrained multivariate clustering analysis tool.  It should be noted that the variables explored may not necessarily accurately predict a child’s learning skills.  Statistical analyses may not necessarily reflect real-life situations and individual behavior.

Lab 3: Crime Analysis (Jan. ’05 – Mar. ’06, Ottawa-Nepean)

Using CrimeStat, crime data (including residential break and enter crimes, commercial break and enter crime, stolen vehicles and robberies) were examined with nearest neighbour analyses, Knox analysis, Moran’s I statistic, hot spot analysis, nearest neighbour hierarchical spatial clustering and kernel density estimation (both single and dual surfaces).  See Crime_Elizabeth_Lee.