Population Change and It's Effects on Redistricting in Michigan

This was a map created for my Political Geography class during the Fall of 2012. The idea of this portion of the project was to identify how population change in the state of Michigan affected the redistricting of congressional districts. 

The map on the far left depicts population change by county from the year 2000 to the year 2010. Counties in red have experienced declines in population while counties in green have experienced population growth.

Upon understanding where these areas of population have changed one can then view the maps on the right and understand how these changes in population affect redistricting. Almost all of the congressional districts between 2000 and 2010 had their boundaries changed dramatically. These changes in boundaries greatly affected the legislative breakdown of United States Congresspersons hailing from the state. The breakdown changed dramatically by not changing at all. Despite seeing a dramatic shift in population in the form of a "white flight" out of the Detroit area, not one district changed what political party represented their district. This has become all too prevalent in today's political culture, as the process of Gerrymandering districts has resulted in stagnation of the democratic process. One would expect radical changes in the political breakdown of Michigan's congressional representatives but truthfully nothing changed, despite the massive face lift on Michigan's districts. 

Site Location Analysis: Indoor Golf Facility

These are two maps I created for a MBA student while at the University of Wisconsin Eau Claire. He is opening up a new indoor virtual golf facility and as part of the report he had to give, commissioned me to create a few maps for him to display potential locations. 

The map on the left displays the best potential locations for the new facility, giving greater weight to the facility being near a golf course (lime green polygons). Red areas are deemed "acceptable", yellow areas are "good", while green areas are "optimal" locations. Link to PDF

The map on the right displays the best potential locations for the new facility, giving greater weight to the facility being within a close proximity to Oakwood Mall, a high traffic area. The same color scheme applies to this map as was used in the map on the left. Link to PDF

Both maps took into account multiple criteria (proximity to roads, golf courses and the mall, and also being located near high income and high density urban areas). For the purpose of this map, we assumed that those viewing and using this map had previous knowledge of the city, and thus knew where the high density urban areas of the city were located. This being written, the only criteria that went into the actual mapping process were the proximity to roads, courses, and the mall. A buffer was created on all these data layers and they were then overlaid based on a hierarchy weighing certain variables (Proximity to the mall and proximity to the courses) higher in each separate study. The multiple ring buffers were then calculated from the values obtained from the previous overlay, giving us the final result.

Fire Severity Mapping, USGS Test Site: Stansbury

This was a map developed during my Internship with the United States Geological Survey during the Summer of 2012. This map was developed using a batch toolset written using python scripting. 

The workflow to create this particular map was a 5 step process as follows

Step 1: Obtain DEM of the test site and digitize boundary where control fire was located.

Steps 2, 3, and 4: Run the batched toolsets in order to calculate pre-fire and post-fire Normalized Burn Ratios (NBR), and then subsequently create a dNBR or the difference between the pre/post fire heat imagery.

Step 5: Join the Land Cover and dNBR imagery in order to create the map seen here. Each land cover was assigned a specific value based on the expected propensity of that coverage to burn in a wildfire. The dNBR image was also reclassified in a similar manner, and then a python scripted equation then calculated the final "Burn Severity Index" value. The values ranged from 1 to 7 (Light Green=1, Red=7) which were interpreted as how severe the landscape had changed pre and post fire.

More about the Monitoring Trends In Burn Severity (MTBS) here: MTBS

Hawaiian Healthcare: An Accessibility Analysis

This was a project the was completed in the Spring of 2012 in my advanced principles of GIS class. The purpose of this project was to identify areas on the Hawaiian island of Oahu that had limited/no access to proper healthcare. More specifically, areas with high Native Hawaiian and Pacific Islander population were also mapped and identified in order to determine whether or not the healthcare deserts were found in the same areas as the aforementioned populations.

In order to determine the "health care deserts" a variety of network analyst techniques were used. The first was a Service Area Network Analysis. To perform this analysis I first determined how large of an area each hospital/clinic should cover by taking the total area of the island and then dividing that by the number of facilities. This provided me with a reasonable assessment of how big of an area each facility should cover. This was step one in determining the healthcare deserts.

I then used another analysis to determine the areas of the island that had a distinctive lack of infrastructure, and thus poor access to healthcare facilities. To perform this analysis I started by geocoding all the healthcare facilities on the island. I then ran the create centroid tool on my census block data layer in order to create a usable reference point to be used in the network analysis. The closest facility analysis was then run and the centroids were connected with the facilities that they were closest to spatially. I then graphed the amount of centroids that each hospital connected to, and found a great disparity in the amount of blocks served. As one would expect, some of the larger hospitals in higher population areas serviced more census blocks, but 2 clinics did not service a single block. 

Perhaps studies such as this could be used in the future to determine how to re-allocate medical resources or propose new infrastructure adjustments that would allow for greater accessibility to medical facilities. I believe this study could be replicated for many of the biggest cities across the United States and similar outcomes could be achieved.   

Link to high quality PDF