Module 1.1 - Calculating Metrics for Spatial Data Quality
For this laboratory assignment, I was tasked with determining the horizontal and vertical precision and accuracy of waypoints to a reference point. I used the batch version of the Project tool to place the given waypoints on the same projection and coordinate field. I then created a multi-ring buffer that was useful in separating the percentiles used for analysis. The analyses were conducted by looking at both the data and running a variety of statistic tools on the data.
Horizontal accuracy consists of how close a point is to the "correct" or reference value. In this scenario, it was determined by measuring the distance between the reference point and the average waypoint. Horizontal precision, on the other hand, is based on how close values are to each other. In this instance, this was determined by looking at how many points lie within the different buffers. The main difference between these two measures, is accuracy looks at the correctness of a value, while precision looks at closeness to other values.
 |
| Horizontal Precision Analysis through Waypoint Buffers |
Horizontal accuracy consists of how close a point is to the "correct" or reference value. In this scenario, it was determined by measuring the distance between the reference point and the average waypoint. Horizontal precision, on the other hand, is based on how close values are to each other. In this instance, this was determined by looking at how many points lie within the different buffers. The main difference between these two measures, is accuracy looks at the correctness of a value, while precision looks at closeness to other values.
The following values were the metrics determined from the provided data:
Minimum: 0.14
Maximum: 6.95
Mean: 2.67
Median: 2.45
RMSE: 3.06
68th percentile: 3.18
90th percentile: 4.67
95th percentile: 5.69
The estimate of horizontal precision of 4.466187 meters was relatively close to the distance between the average waypoint location and the reference point (3.23 meters). This indicates that the results collected using the GPS unit are fairly accurate, as their average is not excessively off from the reference point, only being 3.23 meters away. I found the horizontal accuracy to be approximately 3.23 meters. This means that the calculated horizontal accuracy and horizontal precision (4.466187 meters) are relatively close in value. I calculated the vertical accuracy to be 5.96 meters, being the difference in elevation of the reference point to the average waypoint. This being a relatively low value means that the value was fairly accurate, similar to the results of the vertical precision.
I created a cumulative distribution function graph that shows the relationship between the errors and how much of the error percentage that value consists of. Or in other terms, this graph indicates the distance proportions from the reference point. The Error XY and Cumulative Percentage variables have a strong, positive relationship that appears exponential in nature. All of the metrics calculated in Deliverable 5 could be determined from the CDF, because the graph shows each of the relevant points. Thus, the CDF does provide useful insights that the metrics cannot, by enhancing the data and allowing one to visualize its distribution. Numbers cannot portray certain trends that graphs such as this CDF does.
Minimum: 0.14
Maximum: 6.95
Mean: 2.67
Median: 2.45
RMSE: 3.06
68th percentile: 3.18
90th percentile: 4.67
95th percentile: 5.69
The estimate of horizontal precision of 4.466187 meters was relatively close to the distance between the average waypoint location and the reference point (3.23 meters). This indicates that the results collected using the GPS unit are fairly accurate, as their average is not excessively off from the reference point, only being 3.23 meters away. I found the horizontal accuracy to be approximately 3.23 meters. This means that the calculated horizontal accuracy and horizontal precision (4.466187 meters) are relatively close in value. I calculated the vertical accuracy to be 5.96 meters, being the difference in elevation of the reference point to the average waypoint. This being a relatively low value means that the value was fairly accurate, similar to the results of the vertical precision.
After further analysis, I found there was slight bias in my analysis, as there was potential for user error when using the measure tool in ArcGIS Pro to measure the distance between the reference point and average waypoint. This could be due to not aligning the tool with the actual points or misreading the given measurement. Otherwise, there was very little room for bias to be introduced to this analysis.
CDF Graph