GIS4930 - Scale Effect and Spatial Data Aggregation

 

In this lab I got to see how much the scale and resolution of data can change the story a map tells.

For the vector part, I compared hydro layers for Wake County at three map scales: 1:1,200, 1:24,000, and 1:100,000. At the large scale (1:1,200) there were tons of tiny streams and waterbodies, with very high total line length, polygon count, perimeter, and area. As I moved to smaller scales, the total length and number of features dropped off a lot. Many small ponds and short stream segments simply disappeared, and the surviving features were more generalized with smoother outlines. So even though all three layers were “Wake County hydro,” the geometric properties were very different just because of scale and generalization.

High Resolution 2 meter cell size (top) vs Low Resolution 90 meter cell size (bottom)

For the raster part, I resampled a 1 meter LIDAR DEM to coarser cell sizes (2, 5, 10, 30, 50, 90 m), ran slope on each one, and then made a scatterplot of cell size vs average slope. The pattern was really clear: as cell size increased, the mean slope went down. The 1 meter DEM kept all the little bumps and sharp breaks in the terrain, so the slope values were higher. Coarser grids smoothed everything out by averaging elevation over larger areas, which made the landscape look flatter on paper. That was a good reminder that terrain derivatives like slope are very sensitive to raster resolution.

In Part 2, I looked at the relationship between percent non-white population and percent below poverty using Ordinary Least Squares (OLS) for different units: block groups, ZIP codes, housing districts, and counties. The slope, intercept, and R² values changed each time I switched the geography, even though it was the same people and the same variables. That is basically the Modifiable Areal Unit Problem (MAUP) in action. How you draw boundaries and aggregate data can change the strength and even the appearance of the relationship.

Worst offender for "compactness"

Finally, I looked at gerrymandering, which is when political district boundaries are drawn to give one party or group an advantage. One way to flag suspicious districts is to measure how “compact” they are. I used the Polsby–Popper score, which is $4πA / P^2$, based on each district’s area and perimeter. I calculated area and perimeter with the Calculate Geometry Attributes tool, added a PP field, and then computed the scores. A value close to 1 means the district is pretty compact. Values near 0 mean the shape is long, skinny, or very irregular. When I sorted by PP score, the lowest values belonged to districts like NC Congressional District 12, MD-3, and FL-5, which have very stretched or chopped up shapes compared to a more “normal” district. I used NC-12 as my screenshot example of a district that clearly fails the compactness test.

GIS4930 - Surfaces - Interpolation

For this lab, I used several interpolation methods to model Biochemical Oxygen Demand (BOD) concentrations across Tampa Bay. Each method took a slightly different approach to estimating values between our 41 sample points.

The Thiessen method divided the bay into zones where each area takes the value of its nearest sample point. It’s simple and easy to visualize but produces blocky, unrealistic boundaries. The IDW (Inverse Distance Weighting) method improved on this by blending nearby values smoothly, assuming that locations closer together are more alike. It gave a natural-looking surface that still stayed within realistic ranges. The Spline methods (Regularized and Tension) created even smoother surfaces, but they sometimes exaggerated values in areas with unevenly spaced samples or outliers.

Overall, these techniques showed how interpolation can turn discrete sample data into continuous surfaces that help visualize water quality patterns. For this dataset, IDW provided the most balanced and realistic result, clearly showing how BOD concentrations vary throughout Tampa Bay.

GIS4930 - Surfaces - TINs & DIMs

 

Comparison of elevation contours derived from a Triangulated Irregular Network (TIN) and a Digital Elevation Model (DEM) 

In this lab we explored how terrain is modeled and visualized in ArcGIS Pro using both TINs and DEMs. We used a TIN as the ground surface in a Local Scene, applied vertical exaggeration, and draped imagery to highlight relief with scene lighting. From a DEM we derived slope and aspect, reclassified elevation/slope/aspect, and combined them into a ski-run suitability map (weights: 40% elevation, 35% slope, 25% aspect). We then compared contours from the TIN to contours from a Spline-based DEM at the same interval: DEM contours appeared smoother, while TIN contours showed more faceting along triangle edges. Overall, the exercise showed how data model choice affects downstream products and interpretation, and why unit settings (Z-factor), exaggeration, and lighting matter for clear 3D communication.

GIS4930 - Data Accuracy - Assessment of Road Network Data Completeness

In this lab, we examined road network completeness in Jackson County, Oregon, by comparing the locally maintained Street Centerlines to the US Census TIGER/Line (2000). The goal of the accuracy assessment was simple: measure how fully each dataset captures the real street network, not by point accuracy, but by total road length.

I projected TIGER to the centerlines’ projected CRS, then intersected each road layer with a 5×5 km grid so only the parts inside each cell counted and any segments crossing cell boundaries were split. After intersecting (when geometry changes), I recalculated segment lengths in kilometers, summarized the totals by grid cell for each dataset, and joined those totals back to the grid. To compare them, I computed a percent difference using Centerlines as the base: $\% \text{Diff} = \frac{\text{Centerlines} - \text{TIGER}}{\text{Centerlines}} \times 100$. Positive values mean Centerlines are more complete; negative values mean TIGER is more complete.

In short, a simple length-based metric provided a transparent way to evaluate completeness: TIGER was more complete overall by total length, but the grid map revealed many local areas where the county centerlines exceeded TIGER.

GIS4930 - Data Quality - Data Standards

Map view in ArcGIS Pro showing the Albuquerque City Street Dataset and the 20 digitized test points, placed at well-defined street intersections for use in the positional accuracy analysis.

We continued our journey into data standards this week with Lab 1.2, building directly on concepts of precision and accuracy from last week’s work. Instead of focusing on GPS waypoints, this exercise shifted our attention to testing the horizontal positional accuracy of two full street datasets: a locally maintained city street layer and the broader national StreetMapUSA dataset.

High-resolution orthophotos served as the independent reference data for this project. I digitized 20 control points directly on the imagery, carefully following NSSDA guidelines to ensure they were evenly distributed across the study area and spaced far enough apart to avoid clustering.

Because the orthophotos were massive and demanding on computer performance, I used a mosaic dataset, a skill I learned in the Applications in GIS course at UWF. Building the mosaic allowed me to seamlessly manage and display the imagery, drawing only what was needed in the map window. This not only kept ArcGIS Pro running smoothly but also made the digitizing process far more efficient and precise.

Once the reference points were complete, I snapped matching points to the centerlines in both the city and StreetMapUSA datasets. From there, I exported all three datasets to Excel, where I calculated coordinate differences (ΔX, ΔY), root mean square error (RMSE), and ultimately converted those values into Accuracy_r(95%), the standard NSSDA measure of positional accuracy.

---

Results

The results highlighted a striking difference between the two datasets:

City Dataset (Alb):
Tested 16.94 feet horizontal accuracy at 95% confidence level (NSSDA).

StreetMapUSA Dataset:
Tested 409.01 feet horizontal accuracy at 95% confidence level (NSSDA).

The city’s dataset proved to be highly reliable, with 95% of features falling within roughly 17 feet of their true location. In contrast, StreetMapUSA’s much larger error of over 400 feet made it unsuitable for parcel-level or detailed mapping applications.

GIS4930 - Data Quality - Fundamentals

We kicked off the Special Topics in GIS course with a lab focused on data accuracy and precision, two important concepts for evaluating GPS data quality. In this exercise, we worked with a set of 50 waypoints collected using a handheld GPS unit. The goal was to figure out how tightly the points clustered together (precision) and how close they were to a known reference point (accuracy). Using ArcGIS Pro, we calculated both horizontal and vertical precision, created buffer zones to visualize the spread of the data, and compared our results to the reference point. To wrap up, we explored error metrics and built a cumulative distribution function (CDF) to get a clearer picture of how GPS errors are distributed across the dataset.

---

Horizontal Accuracy: 3.24 m
Horizontal Precision (68%): 4.5 m

Horizontal precision is a measure of how closely repeated GPS measurements align with one another. In this lab, it was calculated as the distance within which 68% of the collected waypoints fall.

Horizontal accuracy measures how close the average GPS location is to a known reference point. This was determined by comparing the average waypoint location to a surveyed reference point.

In simple terms:

• Precision = consistency (how tightly grouped the points are)

• Accuracy = closeness to the “true” location

---

After finishing our GIS work in ArcGIS Pro, we switched to Excel to calculate detailed error metrics for the GPS points. Using formulas, we compared each collected point to a benchmark location and calculated the distance error for every point. From these values, we determined statistics like the minimum, maximum, mean, and percentiles.

One key metric we calculated was the Root Mean Square Error (RMSE), which is a single number representing the typical error across the whole dataset. In this lab, our RMSE was 3.06 meters, meaning that, on average, the GPS points were about three meters away from the benchmark location. RMSE is especially useful for quickly evaluating and comparing the overall quality of GPS data.

---

The Cumulative Distribution Function (CDF) scatter plot can be used to visually determine data metrics, such as the minimum, median, and maximum error, as well as percentiles (68%, 90%, and 95%). However, it cannot be used to directly determine the mean or RMSE.

The CDF provides a clearer picture of the dataset than individual metrics alone, as it shows how all the errors are distributed. It helps identify whether most GPS points are clustered together or if there are a few outliers with significantly higher errors, extending the line out into a “tail.” This insight goes beyond summary numbers and helps in evaluating the overall quality of the GPS data.