Spatial Data Standards

Testing For Spatial Data Standards

This week's lab continued exploring data quality by completing an analysis of horizontal accuracy for two different datasets from 2006. Two different shapefiles were delivered to us: one from the city of Albuquerque itself and the other from TeleAtlas' StreetMap USA. We were also given 6x6 orthophotos of the study area.

Initially, twenty reference sites were chosen using ArcGIS's geoprocessing tool Create Random Points with a minimum distance of 5,000 feet between each site. These reference points were then manually adjusted to the nearest "ideal" street intersection as seen on the orthophotos. Ideal locations contained clear, simple intersections with 90-degree intersections. Street intersections provide precise and unambiguous locations that can be easily identified. They are also consistent over time and can be repeatedly located with high precision.

Next, test points were created for each dataset along the provided polylines. All points were assigned joining IDs and XY coordinates. Lastly, the test points were compared against the reference set using statistical analysis standards provided by the National Standard for Spatial Data Accuracy (NSSDA).

One outlier was discovered in the StreetMap USA dataset and was removed. The results are as follows:

City of Albuquerque Using the National Standard for Spatial Data Accuracy, the data set tested 18.386 feet (5.604 meters) horizontal accuracy at 95% confidence level.

StreetMap USA Dataset

Using the National Standard for Spatial Data Accuracy, the data set tested 271.820 feet (82.850 meters) horizontal accuracy at 95% confidence level.