Geo Leah

In this week's lab, we practiced identifying features in aerial photos based on properties such as texture, tone, size, shape, shadows, pattern, and association. We also compared the colors of features from a true color photo and a false color infrared photo. For the first photograph, I chose areas that represented five levels of texture (from very coarse to very fine) and tone (from very light to very dark), based on the range within this photograph. For the second photograph, I identified and labeled features based on their shape and size, shadows, pattern, or association/context. For example, I identified the parking lot based on the pattern of lines that make up the parking spaces, and I identified sand dunes based on their association with the ocean and the surrounding landscape.

In our final lab we explored the effects of scale on vector and raster data, the Modifiable Areal Unit Problem, and compactness of gerrymandered congressional districts. For vector data, data at a larger scale is more detailed, resulting in longer polylines and polygons with larger perimeters and areas. For raster data (specifically DEMs), higher resolutions result in higher average slopes. Higher resolution rasters have more cells, meaning they have more data points and capture more variation, which results in higher slopes in the case of elevation. District 12 in North Carolina We then explored the topic of gerrymandering, which is the creation of intentionally distorted political districts to favor one party by splitting constituencies between districts or concentrating them in certain districts. The Polsby-Popper score is a way to measure gerrymandering by measuring how compact a district is on a scale of 0 to 1, in which 1 is the most compact. The formula for the Polsby-Popper sco...

In this week's lab we explored different interpolation methods for water quality in Tampa Bay. Given a set of points measuring Biochemical Oxygen Demand (BOD), we interpolated the surface using Thiessen interpolation, Inverse Distance Weighting (IDW), and both regularized and tension Spline methods. Thiessen interpolation generated a surface of discrete polygons of varying sizes, each with the value of the nearest sample point. IDW generated a smooth surface based on nearby sample points weighted by proximity. The spline methods interpolated along smooth curves based on calculated polynomial functions. Both the spline methods were heavily influenced by several outlier points in the data, which caused them to predict some negative values and some unrealistically high values for the BOD surface. After removing these outliers from the original dataset, the spline methods improved, although the regularized spline method still had negative values. IDW interpolation of BOD in Tampa Bay, ...