Module 6 - Scale Effect and Spatial Data Aggregation

 In this lab, we were tasked with understanding the effects of scale on vector data and resolution on raster data, understanding the effect of the modifiable area-unit problem (MAUP) using OLS analysis, identifying multipart features, and being able to measure the compactness of features. 

The first thing I want to discuss is the scale effect on vector data. The smaller the scale, the more detailed the vector data will be. Specifically in the lab, we were given six data sets of hydrographic features in Wake County at varying degrees of scale: 1:1200, 1:24000, and 1:100000. By comparing each dataset, it was clear that the 1:1200 scale had more detail of the hydrographic features within the county compared to the 1:100000 scale. The same goes for the resolution of raster data. The smaller the cell size, the more detailed the raster will be. We compared the resolution of a DEM at 1 meter with that of 90 meters of resolution. What was observed is that at the 1-meter resolution, significantly more features within the DEM were visible than what was visible with the 90-meter resolution.

Another part of the lab we explored is the problem that is gerrymandering. If you don't know what gerrymandering is, it is the manipulation of voting districts to favor a certain party more than another party. The best way it can be measured is by visually looking at it or by calculating the compactness of a district by calculating the district's Polsby-Popper score. To calculate the score, I utilized this formula:

The closer a district is to the number 1, the more compact a district is. The lower the compactness score, the worse the district is gerrymandered. The worst offending district in the country, with a low compactness score, is Congressional District 12 in North Carolina, with a score of 0.03:

Module 5 - Surface Interpolation

 In this week's lab, we were tasked with carrying out different surface interpolation techniques in GIS, critically interpreting the results from the interpolation techniques, and then comparing and contrasting the different interpolation techniques. Those surface interpolation techniques we explored are Theissen interpolation, Inverse Distance Weighting (IDW) interpolation, and Spline interpolation (Regularized and Tension). The area we created the surface interpolation for was Tampa Bay water quality, specifically, Biological Oxygen Demand (BOD). Each method of interpolation conducted resulted in a different output for the same data. 

IDW is a method of surface interpolation that estimates cell values by averaging out all of the data from the sample points, resulting in this output:

Spline is a method of interpolation that estimates values using a mathematical function to minimize surface curvature, resulting in a smooth surface that directly passes and connects points together. However, there are two methods of spline, regularized and tension. Regularized creates the smooth surface common with splines by changing surface values that may lie outside the set data range set by the collected points. Here is what a regularized output may look like:

A tension spline primarily controls the stiffness of a surface. This results in a less smooth surface that is constrained to the set data range, resulting in this output:

The last method is the Theissen Method, which is the simplest form of interpolation. It assigns each cell location the same value as the nearest point, resulting in precise measurements for the cells. It is done by first converting the points to Thiessen polygons, then converting those polygons to a raster by utilizing the feature to raster tool, resulting in this output:

Overall, this lab definitely opened my eyes to how complex GIS can be, and I learned a lot about different methods of interpolation I can now use at my place of employment.

Module 4 - Surfaces - TINs and DEMs

 In this lab, we were tasked with creating 3D visualizations of elevation models, creating and modifying a TIN using various datasets, and then comparing a TIN and DEM elevation model to each other based on their properties and derivatives. Another aspect of this lab was to explore contour creation utilizing a TIN and DEM. When creating contours with a TIN, the contours appear more angular, allowing for a straight-line-like contour that gives a uniform look. The DEM contours gave a more natural appearance. The contours appeared to conform to the natural terrain and showcase a downward slope, whereas the TIN contours did not. The reason the DEM contours appear as such is that the DEM is made up of uniform grid cells that store elevation data. The more cells you have, the more detail the DEM, which allows for more precise contours.

Here are the two different elevation models:

DEM

TIN

Module 3 - Data Quality - Assessment

 The goal of the accuracy assessment in this lab was to determine the completeness of road networks within Jackson County, Oregon. To conduct the analysis, the TIGER and Centerline roadways needed to be limited to the grids that were provided to us, which almost equally make up the entire extent of the county, with some outliers along the northern boundary. To do that, I utilized the modify features pane, specifically the clip tool and selecting the contain all lines option to clip out any part of the lines outside the grids. The next thing that needed to be done was to limit each roadway to a grid, even if it extends through multiple grids. This was done by utilizing the Split Features tool within the geoprocessing pane. This tool would split any lines into a new piece of data if the line transected across multiple grids. While these lines are limited to a grid, the appearance of the lines as one line is still present. Once the lines were isolated to their individual grids, it was now necessary to get the length sum for each roadway type in a grid. I utilized the summarize within tool in the geoprocessing pane to summarize the sum of each roadway in a grid. This allowed me to find the sum of both TIGER Roads and Centerline Roads to find the percent difference in each grid. To find the percent difference, I added the sums in Excel and utilized this formula: % 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 = (𝑡𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑐𝑒𝑛𝑡𝑒𝑟𝑙𝑖𝑛𝑒𝑠 − 𝑡𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑇𝐼𝐺𝐸𝑅 𝑅𝑜𝑎𝑑𝑠) / (𝑡𝑜𝑡𝑎𝑙 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑐𝑒𝑛𝑡𝑒𝑟𝑙𝑖𝑛𝑒𝑠) × 100%

Once I found the percent difference in each grid, I mapped out the differences in this choropleth map below:

Module 2 - Data Quality Standards

 The purpose of this lab was to determine the quality of road networks and to determine the positional accuracy of two road networks by comparison. To be able to determine the quality and positional accuracy of the road networks, we needed to understand and employ the methodology of procedures set by the National Standard for Spatial Data Accuracy (NSSDA). 

To begin the analysis, 60 test points in total were added to a project area in equal distribution. Twenty test points were placed on ABQ intersections, and 20 test points were placed on USA intersections immediately adjacent to the ABQ intersections. Then, 20 reference points were placed in the immediate vicinity of those intersections to get the most accurate measurement of the actual intersection. Here is what this looked like:

 Then to confirm the accuracy of the ABQ points and USA points, we needed to add point X and Y coordinates to each point to get the coordinates to calculate the accuracy. To begin calculating the accuracy, we needed to export the table to Excel, and once added to Excel, I formatted my table to find these calculations:

Diff in X is X(ABQ) - X(Ref), then Diff in X^2 is the value gotten from subtracting squared; the same steps were done for Y. Then to be able to calculate the sum, average, Root Mean Square Error (RMSE), and the actual NSSDA accuracy, you needed to add the Diff in X^2 and Diff in Y^2 values together. To get the average, I took the sum and then divided by 20, and then to get the RMSE, I square rooted the average, and then to get the final NSSDA accuracy, I multiplied the RMSE by 1.7308 to get the horizontal accuracy. 

To adhere to the NSSDA standards, here is my formal NSSDA accuracy statement:

Using the National Standard for Spatial Data Accuracy, the ABQ data set tested at 24.16 ft or 7.36 meters horizontal accuracy at 95% confidence level.

Using the National Standard for Spatial Data Accuracy, the USA data set tested at 201.52 ft or 61.42 meters horizontal accuracy at 95% confidence level.

Module 1 - Data Accuracy Fundamentals

 

The purpose of this lab was to understand the difference between precision and accuracy, be able to calculate vertical and horizontal position accuracy and precision, and then calculate root-mean-square error and cumulative distribution function. The horizontal precision identified is around 4.16 meters. The horizontal accuracy is difficult to achieve as it is unable to be truly perfect. To measure the horizontal accuracy, you would need precise survey equipment or have landmarks that identify the exact location. The most accurate someone can be is around 1cm of accuracy, and that kind of equipment is expensive. Horizontal precision is measured by multiplying the total number of values by the target percentile. If the percentile was a whole number, then I needed to count the numbers from top to bottom until the number was reached, then I got the average of that number and the one after to get the horizontal precision.

Final Project and other maps created

 After eight weeks this class has come to its conclusion. Our last task was to come up with a project idea that is geographic and meaningful. The project I came up with is exploring the possible correlation between hotdog and beer prices within Major League Baseball stadiums and the cost of living index score of the metro area in which these teams reside. The infographic I created is a mixture of bivariate choropleth for the hotdog and beer prices and a choropleth for the cost of living. The report I wrote that goes into detail about my findings can be found here. 

But the gist of it is that after identifying the prices of hotdogs and beers and analyzing how those prices may correlate to the cost of living, it can be concluded that there are more factors outside of just the cost of living that may factor into the prices of these products. Some of those factors may include team sponsorship, distance to distributors, level of income a team brings in each year, and attendance of fans to each game. It can also be concluded that the cost of living has a larger impact on hotdog prices than beer prices, however minuscule it may be the data trend suggests it. For fans who want to acquire the cheapest hotdog, and cheapest beer prices at the same time, and who reside within a relatively low to neutral cost-of-living will want to visit the Arizona Diamondbacks’ Chase Field, the Atlanta Braves Truist Park, and the Minnesota Twins Target Field. Here is the infographic that I created.

The one detail I am pretty proud of is the custom symbology I created for the callouts to showcase the team's location. Now if I ever want to make an MLB-focused map again I have the symbology ready to go. 

Now, I want to showcase some other maps that I created in this class. I am pretty proud of how they turned out. They weren't originally posted because they weren't required for the original blog post of their associated module.

This map is another version of the map shown for the module 1 blog post, the major difference is that this one only depicts the major riverways in Mexico.

Just as the title of the map suggests, this is a map of San Francisco's parks. I used the maps that I created at my job for Sarasota County as a reference on how to showcase parks, highways, and bodies of water in a government format.

This was a fun map to create, as the data was provided but we were allowed to customize how we presented the data as long as it was clear on what are recreational features within the city of Austin. I even went above and beyond and acquired data showcasing the boundary of the city of Austin. By including that boundary it would make it clear to visitors coming to the city what recreational areas are actually within the city limits.

This map was originally a conservation-focused map, but we were tasked with taking this data and formatting the layout to be something that a company would utilize. So, as someone who works for a company, I took inspiration from the layouts that I create and utilize almost every day. 

This map is a simple elevation map of Applegate Oregon. At my job I work with groundwater elevation contours, so by creating this map I gained new insight and ideas on how to present elevation data in a clear manner.