Module 2: Data Quality Standards

The goal of this module was to calculate an accuracy estimate for two similar datasets using the methods described in the the National Standard for Positional Accuracy (1999) Positional Accuracy Handbook. Road centerline data created by Street Maps USA and ABQ Streets were provided and viewed using ArcGIS Pro in order to compare to accurate aerial orthophotos. A point feature class was created for the StreetsMaps USA, ABQStreets, and the reference imagery. Then, 20 random sample points were selected within a study area and a point was placed at the center of the mapped road intersection for each dataset and the actual intersection location according to the aerial imagery. Therefore, each sample location had three points. A screenshot of the 20 sample locations is shown below, in which the red lines are the StreetMaps USA data, the green are the ABQ Streets, and the blue are the reference points based on aerial imagery. 

A screenshot of a sample intersection is also shown below: 

Then, the ADD XY Tool was used to calculate the X and Y coordinate for each point for all three datasets. The tables were then exported to an excel file using the TABLE TO EXCEL tool. Following the accuracy estimate worksheet, the independent points (imagery-based) X and Y coordinate was compared to the test dataset and the NSSDA accuracy, sum of the difference in X and Y, average, and the Root Mean Square Error was calculated. See table below for the StreetMap USA NSSDA value calculation: 

Point number	Point description	x (independent)	x (test)	diff in x	(diff in x) 2	y (independent)	y (test)	diff in y	(diff in y) 2	(diff in x)2 + (diff in y)2
1	SM1	1517725.19	1517606.935	118.25042	13983.1618	1486344.76	1486204.189	140.57321	19760.8273	33743.98903
2	SM2	1523097.74	1523121.269	-23.52948	553.636453	1485862.11	1485796.345	65.761679	4324.59849	4878.234942
3	SM3	1522770.6	1522784.497	-13.900563	193.225645	1482315.61	1482274.676	40.92938	1675.21415	1868.439799
4	SM4	1533213.77	1533265.664	-51.894909	2693.08161	1482951.66	1482980.079	-28.418578	807.615595	3500.697209
5	SM5	1532477.37	1532563.089	-85.719645	7347.85751	1497007.24	1497068.248	-61.007752	3721.9458	11069.80331
6	SM6	1506592.75	1506635.582	-42.835872	1834.91196	1478799.66	1478667.53	132.1267	17457.466	19292.37794
7	SM7	1499494.27	1499571.49	-77.224911	5963.6869	1493712.06	1493712.057	0	0	5963.686905
8	SM8	1522336.01	1522283.026	52.982506	2807.1459	1496833.92	1496820.159	13.759487	189.32348	2996.469379
9	SM9	1504327.04	1504401.264	-74.228198	5509.82538	1487462.36	1487524.498	-62.137999	3861.13093	9370.956308
10	SM10	1515060.27	1515045.331	14.936978	223.113312	1493255.77	1493242.151	13.619067	185.478993	408.5923045
11	SM11	1510697.04	1510711.978	-14.935666	223.074109	1496110.01	1496136.642	-26.632165	709.27219	932.3462992
12	SM12	1505631.65	1505650.77	-19.119384	365.550857	1493215.6	1493210.819	4.7798461	22.8469285	388.3977858
13	SM13	1530454.49	1530539.669	-85.182244	7256.01475	1477870.24	1477896.217	-25.979935	674.957018	7930.971768
14	SM14	1527894.22	1527949.577	-55.354548	3064.12599	1483982.97	1483960.566	22.405467	502.004951	3566.130945
15	SM15	1515332.04	1515234.57	97.474871	9501.35041	1481253.59	1481053.838	199.75485	39902	49403.35044
16	SM16	1497736.46	1497900.517	-164.06004	26915.6965	1478936.16	1478969.112	-32.949409	1085.66356	28001.36007
17	SM17	1500867.09	1500680.775	186.31032	34711.5364	1501038.75	1500746.234	292.52041	85568.1916	120279.7281
18	SM18	1503852.83	1503906.505	-53.671153	2880.59261	1483358.91	1483356.21	2.7030786	7.30663383	2887.899245
19	SM19	1517726.27	1517704.649	21.622988	467.553621	1499856.06	1499887.726	-31.662338	1002.50366	1470.057284
20	SM20	1520651.16	1520611.004	40.156744	1612.56408	1490912.66	1490840.067	72.591062	5269.46231	6882.026382
									sum	314835.5154
									average	15741.77577
									RMSE	125.4662336
									NSSDA	217.1569571

The ABQ Streets NSSDA worksheet table is shown below: 

Point number	Point description	x (independent)	x (test)	diff in x	(diff in x) 2	y (independent)	y (test)	diff in y	(diff in y) 2	(diff in x)2 + (diff in y)2
1	ABQ1	1517725.19	1517730.615	-5.4301073	29.4860648	1486344.76	1486336.919	7.8431602	61.5151614	91.00122617
2	ABQ2	1523097.74	1523089.293	8.4465054	71.3434537	1485862.11	1485854.867	7.2398149	52.41492	123.7583737
3	ABQ3	1522770.6	1522764.418	6.1778092	38.1653261	1482315.61	1482329.892	-14.286389	204.100903	242.2662296
4	ABQ4	1533213.77	1533221.924	-8.1548393	66.5014045	1482951.66	1482941.528	10.132198	102.661428	169.1628324
5	ABQ5	1532477.37	1532475.052	2.3169245	5.36813914	1497007.24	1497012.26	-5.019675	25.197137	30.56527619
6	ABQ6	1506592.75	1506588.523	4.2230887	17.8344779	1478799.66	1478806.293	-6.6364697	44.0427296	61.87720746
7	ABQ7	1499494.27	1499486.543	7.7224255	59.6358556	1493712.06	1493717.463	-5.4058291	29.2229881	88.85884369
8	ABQ8	1522336.01	1522332.687	3.3215157	11.0324663	1496833.92	1496831.863	2.0560983	4.22754002	15.26000636
9	ABQ9	1504327.04	1504326.755	0.2811674	0.07905512	1487462.36	1487451.957	10.403194	108.226454	108.3055093
10	ABQ10	1515060.27	1515057.193	3.0751251	9.45639428	1493255.77	1493253.574	2.1965179	4.82469098	14.28108526
11	ABQ11	1510697.04	1510699.021	-1.9793267	3.91773437	1496110.01	1496104.972	5.0383758	25.3852302	29.30296457
12	ABQ12	1505631.65	1505626.871	4.7798461	22.8469286	1493215.6	1493216.161	-0.5623348	0.31622047	23.16314905
13	ABQ13	1530454.49	1530456.467	-1.9793268	3.91773439	1477870.24	1477867.178	3.059049	9.3577808	13.27551519
14	ABQ14	1527894.22	1527890.268	3.9540603	15.6345931	1483982.97	1483974.186	8.7863998	77.2008206	92.83541373
15	ABQ15	1515332.04	1515333.418	-1.3730288	1.88520795	1481253.59	1481257.712	-4.1187582	16.9641688	18.84937677
16	ABQ16	1497736.46	1497731.652	4.8047804	23.0859148	1478936.16	1478926.552	9.610217	92.3562707	115.4421856
17	ABQ17	1500867.09	1500864.926	2.1591164	4.6617837	1501038.75	1501037.135	1.6194193	2.62251899	7.28430269
18	ABQ18	1503852.83	1503850.131	2.7027505	7.30486027	1483358.91	1483356.21	2.7030786	7.30663383	14.61149409
19	ABQ19	1517726.27	1517727.43	-1.1581342	1.34127475	1499856.06	1499863.014	-6.9501173	48.3041309	49.64540561
20	ABQ20	1520651.16	1520642.28	8.8805597	78.86434	1490912.66	1490926.945	-14.286717	204.110278	282.9746179
									sum	1592.721015
									average	79.63605076
									RMSE	8.923903337
									NSSDA	15.4454919

While creating sample points, I noticed that both datasets were not accurate when compared to the aerial orthophotos road intersection locations. Before making calculations, the Street Maps USA data seemed to be the least accurate of the two. The final step was to define the accuracy of each dataset using a formal horizontal accuracy statement. Both statements are defined below:

• ABQStreets Positional Accuracy Statement: Using the National Standard for Spatial Data Accuracy, the data set tested 15.4 feet horizontal accuracy at 95% confidence interval.

  StreetMaps USA Positional Accuracy Statement: Using the National Standard for Spatial Data Accuracy, the data set tested 217.2 feet horizontal accuracy at 95% confidence interval.

In other words, a road intersection point would be within a defined distance from the actual road intersection 95% of the time. Therefore, the final horizontal accuracy value calculated using the NSSDA method confirms the observation that ABQ Streets is more accurate. In fact, the ABQ Streets dataset is approximately 14 times more accurate than the StreetMaps USA road data!