The URL can be used to link to this page

Home My WebLink AboutTRE License Plate Study 2014 ATTACHMENT 2 AA) 20 : 4 i 1N ITVC AIL\ VAY P R E SS LICENSE PLATE STUDY September 16th to 18th, 2014 Covering the following Park-n-Ride Stations: T & P Station, Hurst/Bell Station, Richland Hills Station, Centreport Station, West Irving Station, and Downtown Irving/Heritage Station. ETC Institute in Association with North Central Texas Council of Governments December 31,2014 2. ETC Institute positioned surveyors at each entrance of the Park-n-Ride lots from the hours of 6am to 8pm. Their responsibility was to view and record license plates into the iPads as vehicles exited and entered the lots. 3. From the hours of 6am to 9:30am (AM Peak),surveyors positioned at the entrances and exits of the lots recorded the license plates of vehicles EXITING the lot only; this allowed Kiss-n-Ride vehicles to be captured efficiently and safely. 4. From 9:30am to 8pm,which was classified as after the AM peak,surveyors recorded the plates of vehicles ENTERING the lot only. 5. In addition to surveyors at each entrance/exit, beginning at 6am up to three (3) "floating" surveyors walked from parking space to parking space recording the plates of parked vehicles. Surveyors recorded each plate and marked the back tire of the vehicle they recorded with chalk. The"floating"surveyors were strategically placed so that they could record the entire lot and mark vehicles as efficiently and quickly as possible. The number of floating interviewers varied depending on the size and physical layout of each lot. The days on which each station was surveyed is shown in Exhibit 1. Exhibit 1:Schedule of Stations to be Surveyed Date Stations to be Surveyed Tuesday,September 16,2014 Downtown Irving/Heritage Center,T&P Station Wednesday,September 17, 2014 Centreport Station, Hurst/Bell Station Thursday,September 18, 2014 Richland Hills Station,West Irving Station The maps in Exhibit 2, Exhibit 3, Exhibit 4, Exhibit 5, Exhibit 6, and Exhibit 7 show the position of each surveyor by station for the AM peak (6:00 am to 9:30 am) and after AM peak (9:30 am to 8:00 pm) for Downtown Irving/Heritage Center; T&P Station, Centreport Station; Hurt/Bell Station; Richland Hills Station;and West Irving Station respectively. -2- Exhibit 3:Position of each surveyor for T&P Station for the AM Peak and after the AM Peak T&P Station -Data Collected:Tuesday 9-16-14 -4 Interviewers from 6am to 930am tntrance Surveyor -Parking Spaces=387 ilostor yr.Not :.w t II.7P+4 farameti - . • - . • J _ -` ,IfefiT • I 1••.I tit..)��:. . ii..1 1-',;', ___,„, , r I :1101k1! .r -! �t ! ¢ 11 �etli ,� � .4,___ „, .,intet:00.7.. r_ .. .: ., ' r r�,eft ¢ i• _ . pt p:r r f1T0f• f f t ... .: _., .___......1„,..--_,..,_ ,.... . i_.1.,.,,,v„... ,.,:,, ..., ..., .„ . .LA___!... ,,, , , _ ..._ ...:„..4!.., :.__ ._ . _,___ r - — --- .._ _ _ RI 1 al -_ t�r;,_- r - ~) • / etas i T&P Station -Data Collected:Tuesday9-16-14 -2 Interviewers from 930am to 8pm Eiltrancr•Susvr•yar -Parking Spaces=387 nail*, is rnl Or no 1[ n. - �2`'L.1•!!r! err �HI ' • + rg ER�:��tllr:41;ri :41. ltllli �r� 6etrlFtr��Ft I - L i . a~ • ,C ret fit 'L I ;; [1fffIT , if 111 . r .. , ,.... • . 1 . . 41. . .I -4- 1 Exhibit 5:Position of each surveyor for Hurst/Bell Station for the AM Peak and after the AM Peak Hurst/Bell Station -Data Collected:Wednesday 9-17-14 -4 Interviewers from 6am to 930am trtttence Surveyor -Parking Spaces=488 • finatrr ;, a. 11srr4/rrintrpni,v:tt �irfr�'RM ` �eCKriri�%F1i1?ilpi g • .. :.! '.: � c�afrwEft�rR6uii.r: y..`�• f�;��rrff.raf arapf ;► a 1! x f1lfa 4 a Y Na .� {- A ,n ` .. nig - Hurst/Bell Station -Data Collected:Wednesday 9-17-14 -1 Interviewer from 930am to 8pm ErrlrHrxe Surveyor -Parking Spaces=488 ii finale! {ti t • :rt . _ .. - lifr rK µ f vrit vc tS ly�r YYl��t11 ffc(f tit�tt�r4gfMCr# trt�/!'ryliPrri►�r �•.. i.' .. / tt��c r Ott #w S htgrifirliulSr e ntriraMtatt- I ..,14 ;4a��tfr¢tHi c .4.V$ �IfF r rl ►'f -• i lO;4,4‘:4►,I -6- r Exhibit 7:Position of each surveyor for West Irving Station for the AM Peak and after the AM Peak W Irving Station -Data Collected:Thursday 9-18-14 -4 Interviewers from 6am to 930am entrance Suavnyor -Parking Spaces=500 amok .I. - .� ` ,r a s $ ' t3 � i', • ga. -* irill'4$- V 4lifd••4 4. +r am-^� .may . "'"' -:..v-_ 'mil-A_ .4.' itt 1 W Irving Station -Data Collected:Thursday 9-18-14 -2 Interviewers from 930am to 8p tntrarlccrSurveyor -Parking Spaces=500 '' • rlwtar . s _ `t , ll r - !,. . ,.,44t l' -,.- , • •'— IS ' , rt18 r . Wyk e � g E f ��l ali a f I Il 4 rff;. Ea�11 tittlfffa•SI- y: --Mir I J _ .. •Z. ��. -. . �..• n _ ' r- - - ^ ' �F` . ,'-_r— I. -8- Matched Records The 2,928 license plates that were captured during the three-day survey were checked by North Central Texas Council of Governments (NCTCOG) against the Texas vehicle registration data to obtain the city in which each vehicle was registered. 368 records,which is less than 13%of the plates collected, could not be matched against the registration database. 319 of the 368 unmatched records were from Texas. The main reasons for unmatched records from Texas license plates seem to be errors in capturing the correct license plate characters in the tablets. Exhibit 10 shows the breakdown of the unmatched license plates. Exhibit 10:License Plates which were unable to be matched too city Index State on License Plate It Plates %Plates 1 [Blank] -From TX 319 10.89% 2 [Blank] -Out of State 43 1.47% 3 [Blank]-State not Recorded 6 0.20% TOTAL 368 12.57% City Corrections 218 cities were found among 2,560 matched plates.There were two corrections that needed to be done during this matching process. First,it was found that the registration database contained misspellings of cities. Therefore,the cities for each license plate had to be reviewed for spelling errors before the matched city was finalized. A second problem that was found was that the Zip Codes could represent multiple cities. A Zip Code could geographically cover multiple cities even though it is only assigned to a single city in the United States Postal Service's(USPS)database.Therefore,any analysis solely based on the name of the cities in a list of physical addresses has an inherent reporting error in areas where multiple cities share the same Zip Code. An example of this overlap is Zip Code 76040 that is shared by the cities of Euless and Fort Worth. To eliminate this reporting error, NCTCOG identified the registration city by overlapping the geocoded home street address reported in the vehicle registration database with the city boundary shapefile. For any license plate whose home street address was incomplete or not geocodable,the original city reported in the registration database was used. The resulting changes in numbers in the nine mid-cities is shown in Exhibit 11. -10- Exhibit 13: License Plate Records by Owner City—Top Contributors Index Owner City #Records %Records 1 FORT WORTH 576 22.50% 2 IRVING 318 12.42% 3 ARLINGTON 222 8.67% 4 EULESS 152 5.94% 5 GRAND PRAIRIE 132 5.16% 6 DALLAS 121 4.73% 7 BEDFORD 99 3.87% 8 HURST 99 3.87% 9 NORTH RICHLAND HILLS 72 2.81% 10 KELLER 41 1.60% 11 GRAPEVINE 38 1.48% 12 HALTOM CITY 35 1.37% 13 WATAUGA 23 0.90% 14 RICHLAND HILLS 22 0.86% 15 MANSFIELD 20 0.78% 16 COLLEYVILLE 19 0.74% 17 HOUSTON 18 0.70% 18 ATLANTA 16 0.63% 19 AUSTIN 15 0.59% 20 LEWISVILLE 15 0.59% TOTAL 2,053 80.20% Accuracy of the Results Similar to any other survey,the license plate survey of September 2014 has some inherent limitations that stem from the following: 1) natural variation of observed data on different days, 2) survey design assumptions,and 3)vehicle registration data matching.Limitation of accuracy due to the natural variation of observed data is a normal characteristic of all surveys. By only looking at license plates of cars which use the Kiss-n-Ride and Park-n-Ride, the survey design assumes that TRE passengers from these cities are well-represented by all those who drive or are driven to the stations. Whether this assumption is creating a large bias in the results or not can be examined through comparison against the results of the onboard interview survey of TRE passengers in April 2014, which captured a 10%sample of riders.That survey shows that about 76%of the TRE riders whose home is in the nine mid-cities used drive access mode to TRE stations. Since this percentage is high, it is reasonable to assume that the survey results are accurate enough for this application. Finally, the license plate matching process introduces two potential issues. First, the unmatched plates may be distributed disproportionally among the nine cities.This error is probably insignificant because we matched a high percentage of the plates. The second issue lies in the assumption of accepting the place -12- Lessons Learned Overall the methodology, survey, and implementation plan developed by ETC Institute in cooperation with NCTCOG was extremely effective. One improvement that could be made for future studies of similar size and scope would be positioning an off-duty transit police officer onsite. A very small number of drivers who parked their vehicle expressed concern about the legitimacy of the survey. Even though surveyors were equipped with signed authorization letters from the proper authority, drivers were uneasy and looked for the study to be verified by police authorities. For this reason, ETC Institute recommends hiring an off-duty transit police officer during the AM peak only(6:00 am to 9:30 am),since this is when surveyors are most visible. While the goals of the survey were accomplished, the survey had very minimal lead time before implementation. The results were needed quickly, and the survey had to be completed before the Texas State Fair began since the fair could affect TRE ridership. There was adequate time for the transit agencies to approve and inform their transit policy and security. Still, the transit agencies received various calls regarding concerns about the purpose and organizers of the license plate survey. If conducted again,the survey should allow for more time for the Dallas Area Rapid Transit (DART) and The Fort Worth Transportation Authority(The T)to inform their riders of the survey effort. An issue was also found with the city listed in the Texas vehicle registration database, where a Zip Code could actually be split between two cities. To rectify this issue, NCTCOG corrected the registration database using the geocoded home address and city layer. This issue was discussed in detail in the Matched Records section. It is important that future implementations of the survey similarly make corrections to the registration database based on the city layer. Conclusion The survey method and the accuracy analysis along with comparison of the previous shares show the results of the survey are appropriate for identifying the shares of TRE riders among the nine cities. No significant surprises are found in the comparative analysis of different surveys.The differences are natural and can be explained if needed. -14- Index Owner City #Plates %Plates 36 TULSA 8 0.31% 37 ALEDO 7 0.27% 38 HASLET 7 0.27% 39 RIVER OAKS 7 0.27% 40 SOUTHLAKE 7 0.27% 41 AZLE 6 0.23% 42 DESOTO 6 0.23% 43 LANCASTER 6 0.23% 44 WHITE SETTLEMENT 6 0.23% 45 FARMERS BRANCH 5 0.20% 46 JOSHUA 5 0.20% 47 LIBERTYVILLE 5 0.20% 48 MIDLOTHIAN 5 0.20% 49 ROANOKE 5 0.20% 50 ROWLETT 5 0.20% 51 WAXAHACHIE 5 0.20% 52 DEL RIO 4 0.16% 53 FLOWER MOUND 4 0.16% 54 FOREST HILL 4 0.16% 55 FRISCO 4 0.16% 56 GRANBURY 4 0.16% 57 KATY 4 0.16% 58 MCKINNEY 4 0.16% 59 PANTEGO 4 0.16% 60 AMARILLO 3 0.12% 61 CLEBURNE 3 0.12% 62 CORINTH 3 0.12% 63 DALWORTHINGTON GARD 3 0.12% 64 EL PASO 3 0.12% 65 FORNEY 3 0.12% 66 HUMBLE 3 0.12% 67 KENNEDALE 3 0.12% 68 SPRINGTOWN 3 0.12% 69 VICTORIA 3 0.12% 70 WACO 3 0.12% 71 WICHITA FALLS 3 0.12% 72 WYLIE 3 0.12% 73 ABILENE 2 0.08% 74 ALLEN 2 0.08% 75 ALVARADO 2 0.08% -16- Index Owner City # Plates % Plates 116 BRICK 1 0.04% 117 BROOKLYN 1 0.04% 118 BROWNWOOD 1 0.04% 119 CALDWELL 1 0.04% 120 CANYON CREEK CDP 1 0.04% 121 CARRIZO SPRINGS 1 0.04% 122 CHANDLER 1 0.04% 123 CIBOLO 1 0.04% 124 COCKRELL HILL 1 0.04% 125 CORPUS CHRISTI 1 0.04% 126 CORSICANA 1 0.04% 127 COTULLA 1 0.04% 128 CUMBY 1 0.04% 129 DE CORDOVA BEND 1 0.04% 130 DENISON 1 0.04% 131 DEW APO 1 0.04% 132 DRIPPING SPRINGS 1 0.04% 133 DUBLIN 1 0.04% 134 DUMAS 1 0.04% 135 EAGLE LAKE 1 0.04% 136 EAGLE PASS 1 0.04% 137 EL DORADO HILLS 1 0.04% 138 ELMENDORF 1 0.04% 139 EMORY 1 0.04% 140 FAIRFIELD 1 0.04% 141 FAIRVIEW 1 0.04% 142 FT STOCKTON 1 0.04% 143 GLADEWATER 1 0.04% 144 GLENN HEIGHTS 1 0.04% 145 GONZALES 1 0.04% 146 GRANDVIEW 1 0.04% 147 HALLETTSVILLE 1 0.04% 148 HEATH 1 0.04% 149 HENDERSON 1 0.04% 150 HEREFORD 1 0.04% 151 HICO 1 0.04% 152 HIGHLAND PARK 1 0.04% 153 HILLIARD 1 0.04% 154 HUNTSVILLE 1 0.04% 155 HUTCHINS 1 0.04% -18- Index Owner City #Plates %Plates 196 SANGER 1 0.04% 197 SANTA FE 1 0.04% 198 SCHERTZ 1 0.04% 199 SCHULENBURG 1 0.04% 200 SEAGOVILLE 1 0.04% 201 SEGUIN 1 0.04% 202 SELMA 1 0.04% 203 SHREVEPORT 1 0.04% 204 SMITHVILLE 1 0.04% 205 SOUR LAKE 1 0.04% 206 SPARKS 1 0.04% 207 STAFFORD 1 0.04% 208 TAHOKA 1 0.04% 209 TOLAR 1 0.04% 210 TROY 1 0.04% 211 TUSCOLA 1 0.04% 212 TYLER 1 0.04% 213 UNIVERSITY PARK 1 0.04% 214 WAVE VILLAGE 1 0.04% 215 WEST 1 0.04% 216 WESTWORTH VILLAGE 1 0.04% 217 WILLIS 1 0.04% 218 WILLOW PARK 1 0.04% TOTAL 2,560 100.00% -20-