There and Back Again

In this lab, I worked through a series of exercises designed to compare and apply two fundamental elevation data models: TINs (Triangulated Irregular Networks) and DEMs (Digital Elevation Models). The workflow highlighted how these models can be visualized, analyzed, and compared to better understand terrain. Part A: Draping imagery over terrain This part of the lab had us looking at Death Valley with a remotely sensed image draped over a TIN, which was used as an elevation layer to produce a 3d image. We added vertical exaggeration to reveal surface features that weren't noticeable in the flat imagery. Part B: Ski run suitability modeling Next, I built a suitability model using a DEM. I reclassified elevation, slope, and aspect into categories of suitability, then combined them in a Weighted Overlay. The resulting raster identified areas best suited for ski runs, with greens showing higher suitability and reds showing lower suitability (see screenshot below). I also got to u...

This was a fairly straightforward lab. In it, we had a county boundary, a gridded layer within the county, and two line shapefiles of road data. Our goal was to determine which set of road data was more comprehensive, both overall and within each grid cell provided. To do this, I edited the attribute tables of each to add a length in km field, then used calculate geometry to have the road values in kilometers instead of feet. Then, I exported the tables to Excel and summed the values. From there, I clipped each road layer by the grid layer to remove roads outside the grid, calculated geometry again, and then used the Summarize Within tool to get the length of road inside each cell, then joined them together and used a python expression to determine the percent difference. I probably wasted the most time trying to symbolize it in a map -- there are several large outliers, and several grids with no roads from either dataset. I ultimately decided to mark the null cells as 0/no differ...

This was a fun lab. In it, we started with a two line shapefiles of Santa Fe, NM streets, and a large set of ortho imagery, which we had to add to the map. We then selected 20 roadway intersections on the map, with over 20% per quadrant and meeting other NSSDA standards for sampling (i.e. over 10% of the radius distant from each other). From there, we created point data for the both roadway sets, then found the "correct" point as a reference by looking directly at the ortho images. We then assigned XY values to each point, then used a positional accuracy worksheet from the National Standard for Spatial Data Accuracy to calculate the difference in X and Y values, their squares, the sum of their squares, and finally the sum, average, Root Mean Square Error and the NSSDA accuracy statistic. If you're in this program and are searching out blogs to find examples for this lesson, here are two helpful hints: 1: Your initial orthos will be missing a lot. It's not doc...

This week's lab was all about accuracy and precision. We had a scatter of GPS waypoint coordinates, and had to adapt them to point data for our map. Then we calculated average position of those points, and then calculated how large a radius from the average waypoint would be needed in a series of ringed buffers to capture 50, 68, and 95% of all waypoints. We then used an accurate reference point to determine the relative accuracy and precision of the measurements. Horizontal accuracy is measured by comparing the average GPS location to a surveyed reference point — it tells you how close your measurements are to the true location. Horizontal precision is measured by looking at how tightly repeated GPS points cluster around their average — it tells you how consistent your measurements are with each other, regardless of whether they’re close to the true location. We also made similar calculations for the altitude of the readings, and compared the relative horizontal and vertic...