Module 1.2: Data Quality - Standards

 For this week's module, we utilized the calculations learned previously to find the accuracy and metrics of two different street intersection datasets in comparison to the reference points created based on the orthoimage. Twenty points were created based on visually identifying the street intersection. Those points are depicted below with the two street centerline datasets. 

To create the three datasets for analysis, I picked twenty intersections evenly spaced out in the study area (outlined in black). Once each point was created for the three datasets; ABQ intersection, Street Maps intersection, and reference intersection. I added the XY values and exported each table to a CSV. From there I compared the ABQ and Street Map datasets against the reference points individual. This is where the Sum of Deviation, Average Deviation, and Root-Mean-Square-Deviation (RMSE), and  National Standard for Spatial Data Accuracy (NSSDA) were calculated.

Using the National Standard for Spatial Data Accuracy, the Street Maps dataset tested horizontal accuracy of 275.31ft at a 95% confidence level. 

Using the National Standard for Spatial Data Accuracy, the ABQ dataset tested horizontal accuracy of 43.44ft at a 95% confidence level. 

Based on the accuracy statements and metrics calculated, the ABQ dataset is the most accurate for street intersections because it had the lowest numbers. The lower the number, the better the accuracy. 

As a disclaimer, these measurements have an added measure of error since the reference points were created by visually estimating the center.