Developing a Method Selection Tool for Travel Forecasting (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Final Report Project No. 08-94 41 CHAPTER 3. FINDINGS AND APPLICATIONS 3(A) TRAVEL FORECASTING GUIDE SOFTWARE TOOL The development of the Travel Forecasting Guide (TFGuide) has followed a typical software development process and included several steps: 1. Gathering initial requirements and enumerating a typical user’s needs. 2. Constructing a prototype web application to represent all necessary elements, establishing intuitive interpretations of the technical content, and ensuring that the user/administrative aspects of the tool function properly. 3. Designing and building a GUI to provide a user-friendly presentation (including helpful tips for users) of the technical content. 4. Developing the technical content (since the tool is data driven) for each method and displaying background materials and sources for each method; developing the process to evaluate and weight costs and benefits for the method selection tool. 5. Pilot testing the software with selected transportation agencies. TFGuide is documented in detail in NCHRP Research Report 852: Method Selection for Travel Forecasting: User Guide, which includes a user guide and reference guides (included as appendices) for the technical content in TFGuide. This Web-Only Document provides details on the development of the software tool (Chapter 2), the administrative functionality (B), and the pilot tests completed during the project (3(C)). 3(B) ADMINISTRATIVE FUNCTIONALITY OF TFGUIDE TFGuide was developed to allow administrator-initiated enhancement and refinement. The software can improve over time with additional input from other experts and users with specialized expertise. The administrator can take feedback directly from users or can review user scenarios to evaluate the recommendations for these specific circumstances. The administrative functionality of the method selection software tool comprises eight sections: 1. Methods 2. Method Package Relationships 3. Performance Metrics 4. Programs 5. Requirements 6. Regions 7. Resources 8. Site Users The research team added definitions for methods, programs, requirements, and metrics in the administrative homepage to orient the administrator, as shown in Figure 11. This homepage lists each area, with categories and elements entered separately where necessary.

Final Re FIGURE 11 Method The first methods needed to are revea quickly c port : TFGUIDE— s page within and a summ implement led within e onfirm that ADMINISTRA the Method ary of the bu each metho ach row on the details o TIVE HOME P s portion of dget, sched d (Figure 12 click (shown f each meth 42 AGE the adminis ule, program shows a pa in Figure 1 od are consi trative site i s, requirem rtial list as a 3). This list stent and co Pro s a list of th ents, metric n example) allows the a mprehensiv ject No. 0 e current s, and resou . More detai dministrato e. 8-94 rces ls r to

Final Re FIGURE 12 FIGURE 13 port : PARTIAL A : PARTIAL A DMINISTRATI DMINISTRATI VE LIST OF M VE LIST OF M 43 ETHODS ETHODS WITH DETAILS Project No. 08-94

Final Re The adm page leng the metho budget an identifyin class met administr presented FIGURE 14 Some of travel, su identified appears. port inistrative fo th is too lon d, assignme d schedule, g the indivi hods (shown ators to ens with less-c : PARTIAL S the methods ch as mode more clear rm for addi g to include nt of releva identifying dual resourc in Figure 1 ure users wh apable recom CREENSHOT listed in the choice (fixe ly by the adm ng a new me in its entire nt programs the requirem es needed to 5). The incl o specify su mendation OF ADMINIST Figure 15 a d factors) an inistrator b 44 thod is part ty here. Thi for which i ents and m implement usion of the pporting on s in the men RATIVE FOR re specified d time-of-d y hovering ially present s form inclu t may be use etrics satisfi the method “lower clas e method at u. M FOR NEW M separately ay (fixed fac over the me Pro ed in Figure des naming ful, setting ed by that m , and assign s methods” their agenc ETHOD for passenge tors). These thod so that ject No. 0 14; the web and describ a range for t ethod, ment of low feature allow y will not be r and freigh can be a descriptio 8-94 ing he er s t n

Final Re FIGURE 15 Each met These me than one performa and the n demonstr for a user 1. M 2. M 3. M The adm relationsh defining port : “LOWER C hod is relate thod packag method to s nce metrics uances of ho ation purpo ; the possib ethod 1 can ethod 1 is b ethod 1 req inistrator de ip pairing t new method LASS METHO d to other m e relationsh uccessfully . A tiered rel w various m ses that two le ways thes pair with M est if paired uires Metho fines and ma hat is most a package re DS” ELEMEN ethods to fo ips were ne complete a p ationship st ethods pair methods (M e two metho ethod 2. with Metho d 2. nages the m ppropriate. lationships. 45 T ON THE AD rm method cessary sinc roject and m ructure was with others ethod #1 an ds can be c d 2. ethod pack Figure 16 sh MINISTRATIV packages th e most plann eet certain architected . For examp d Method # ombined inc age relations ows the adm Pro E FORM FOR at are presen ing problem requiremen to support th le, assume f 2) can be pr lude the fol hips based inistrative ject No. 0 A NEW METH ted to the u s rely on m ts and is functiona or esented toge lowing: on the type o form for 8-94 OD ser. ore lity ther f

Final Re FIGURE 16 Program The adm users. Th or more p categorie this inclu methods port : ADMINISTR s inistrator ma ese transpor lanning issu s. The admi des the abili can add valu ATIVE FORM nages the c tation-relate es or questi nistrator can ty to add a d e. FOR NEW ME ategories an d policies, p ons. Curren add or chan escription f 46 THOD PACK d list of plan rograms, pl tly, program ge program or the progr AGE RELATIO ning progra ans, or inve s and plans categories am and the Pro NSHIP ms or plans stment proje include five and element ability to ide ject No. 0 of interest t cts address major s (Figure 17 ntify which 8-94 o one );

Final Re FIGURE 17 The adm demonstr administr port : ADMINISTR inistrative si ate how adm ative functi ATIVE FORM te also inclu inistrators on to add pr FOR NEW PR des an overv can review t ograms is sh 47 OGRAM CAT iew of the p he technical own in Figu EGORY rograms (b content (by re 18. Pro y category a program). T ject No. 0 nd method) he 8-94 to

Final Re FIGURE 18 Require The adm and can a for addin descriptio requirem user may whether r Results c then cont the admin port : ADMINISTR ments inistrator pro dd or refine g a new requ n, identifyi ent category select multi equirement an be saved inue editing istrative for ATIVE FORM vides detail the categor irement cat ng the inclu is addressin ple requirem s can be lim in one of th ; and 3) sav m for a new FOR NEW PL s regarding ies of requir egory. This ded element g. This form ent elemen ited to a sing ree ways: 1) e, and then c requiremen 48 ANNING PRO the requirem ements. Figu form includ s, and identi also includ ts. This flex le response save, and th lose. These t category. GRAM OR PL ents or sco re 19 prese es naming t fying what p es a checkb ibility allow or can acce en add ano options are Pro AN pe of the pla nts the adm he category, lanning qu ox to determ s the admin pt multiple r ther categor included at ject No. 0 nning progr inistrative fo adding a estion this ine whethe istrator to de esponses. y; 2) save, a the bottom o 8-94 am rm r the cide nd f

Final Re FIGURE 19 The requ the indiv administr requirem shown in methods administr identifyin port : ADMINISTR irements are idual level. R ator can add ent is design Figure 20. T once the req ator can als g the releva ATIVE FORM listed for th equiremen requiremen ed to addre he adminis uirements a o add new re nt associate FOR NEW RE e administr t elements c t elements b ss and the m trator can re re programm quirements d methods ( 49 QUIREMENT ator by elem an also be d y identifyin ethods that view the lis ed into the by naming Figure 21). CATEGORY ent, allowin eleted using g the plann are relevant t of requirem method sele and describi Pro g changes to the action m ing question to a require ents with a ction tool. T ng the requi ject No. 0 be manage enu. The that a ment elemen ssociated he rement and 8-94 d at t, as

Final Re FIGURE 20 FIGURE 21 Resourc The adm comparab 22 shows shown pr resource administr resources port : ADMINISTR : ADMINISTR es inistrator de le manner t the form us ovides spac categories a ator form to —with cate ATIVE FORM ATIVE FORM velops techn o what has b ed by admin e for an opti nd the resou add a new gories and b FOR NEW RE FOR NEW RE ical content een present istrators to onal descrip rces that be resource. Th udgets—to 50 QUIREMENT QUIREMENT for the reso ed for meth create a new tion. The ad long in each e administr verify and c ELEMENT urce catego ods, program resource c ministrator category. F ator also has ompare with Pro ries and reso s, and requ ategory; the can also rev igure 23 sho an availabl other resou ject No. 0 urces in a irements. Fi label sectio iew the list ws the e list of rces. 8-94 gure n of

Final Re FIGURE 22 FIGURE 23 Perform TFGuide measures comparab Figure 24 this inclu selecting performa form use connecte review th Figure 26 currently port : ADMINISTR : ADMINISTR ance Mea currently in . Administr le manner t shows the des identify whether mu nce measure d by admini d to a metho e list of per shows the included as ATIVE FORM ATIVE FORM sures cludes six m ators can ad o what was form used b ing relevant ltiple respon categories strators to ad d as either a formance m form used b examples o FOR NEW RE FOR NEW RE ajor catego d or refine th presented fo y administra parameters ses are allo and metrics d a new per direct or in etrics, classi y administra f performan 51 SOURCE CA SOURCE ries of trans e performan r methods, p tors to creat for each per wed. The ad that belong formance m direct produ fied by perf tors to creat ce measure TEGORY portation pla ce measure rograms, re e a new per formance m ministrator in each cate easure elem ct of the me ormance ele e a new per elements, bu Pro nning perfo categories quirements formance m easure categ can review gory. Figure ent. Each e thod. The a ment and ca formance m t not indivi ject No. 0 rmance and metrics , and resourc easure categ ory and the list of 25 shows t lement is dministrator tegory. etric. Metric dually conne 8-94 in a es. ory; he can s are cted

Final Re to metho The adm methods— FIGURE 24 port ds or parame inistrator can to verify a : ADMINISTR ters, since t also review nd compare ATIVE FORM here are hun the list of with other m FOR NEW PE 52 dreds of dif performance etrics. RFORMANCE ferent metri measure el CATEGORY Pro cs that could ements—wi ject No. 0 be include th metrics a 8-94 d. nd

Final Re FIGURE 25 Special N port : ADMINISTR ote: Content ATIVE FORM for demonstr FOR NEW PE ation purpos 53 RFORMANCE es only. ELEMENT Project No. 0 8-94

Final Re FIGURE 26 3(C) P Pilot test  As  Re  Vi  Du The feed software agencies participat Julie Dun perspecti pilot test Associa AMBAG Overall, A worked w FIGURE 27 TFGUIDE port : ADMINISTR ILOT TES s were cond sociation of gional Tran rginia Depa nbar Transp back receive tool. Origin agreed to co e due to wo bar to partic ve on how t participant a tion of Mo tested eight MBAG sta ell and was : AMBAG SC ATIVE FORM TS ucted with f Monterey B sportation C rtment of Tr ortation Co d as part of ally, four ag mplete the p rkload conc ipate on be he tool may nd provided nterey Ba scenarios, ff thought t intuitive. ENARIOS FO FOR NEW PE our agencies ay Area Go ommission ansportation nsulting (DT these pilot t encies were ilot tests, b erns. Instead half of the T be used. RS backgroun y Area Go as shown in he software R PILOT TEST 54 RFORMANCE and organi vernments (RTC) (VDOT) C) ests helped contacted to ut FDOT, D of identify FResource i G also cond d materials vernments Figure 27. tool ING METRIC zations in O (AMBAG) improve the participate istrict 5 in O ing an altern nitiative an ucted indivi on the devel “W gr wh Pro ctober 2016 clarity and in the pilot rlando was ative agenc d provide a n dual webina opment of t e think the to eat and perfo ile being eas ject No. 0 : usefulness o tests; three unable to y, RSG aske onagency rs with each he tool. ol looks rms well y to use.” 8-94 f the d

Final Re AMBAG notes or c staff mem team add environm the test d This has be used t selected m prompted methods. port staff provid omments o bers about ed this addit ental justice ue to limitat been correct o support ev ost of the p many meth ed feedback n the home s the significa ional eleme scenario th ions on whi ed in the too aluation of erformance od recomm after testin creen to all nt details of nt on the ho at AMBAG ch methods l to reflect environment metrics and endations fo 55 g. AMBAG ow internal each scena me screen, a tested did n could suppo the fact that al justice pr a high leve r advanced staff recom communicat rio. Based o s shown in ot return an rt the enviro most passen ograms (Fig l of spatial d and more de Pro mended add ion among m n this feedb Figure 27. T y recommen nmental jus ger-oriente ure 28). Th etail (micro tailed simul ject No. 0 ing an area anagers an ack, the rese he dations dur tice program d methods c is scenario a zones), whi ation-based 8-94 for d arch ing . an lso ch

Final Re FIGURE 28 Regiona RTC test that seve alphabeti alphabeti were ord period (1 methods, required FIGURE 29 RTC staf evaluatin scenario. scenario input the scenario. port : RECOMME l Transpo ed two scen ral itemized cally. For ex cally (Figur ered starting 5 minutes o and resourc to identify th : ORIGINAL A f also provid g several sc Based on th with the cur se details on NDATIONS FO rtation Co arios using T lists could b ample, the e 29), but R with the lon r less). All li es are now e sort order ND NEW TEM ed input mo enarios, it w is feedback rent method ce. These ca R ENVIRONM mmission FGuide. RT e made clea desired leve TC thought gest time p sts that are ordered logi for any new PORAL DET dification s as necessary , the researc s and resour n then be ch 56 ENTAL JUST C staff prov rer if they w l of tempora this informa eriod (annua currently pre cally rather elements a AIL ORDER uggestions. to re-input h team revis ces from the anged or ad ICE SCENAR ided feedba ere ordered l detail was tion would b l) and endin sented in re than alphab dded to the Originally, w current met ed TFGuide prior scena apted as ne Pro IO FOR AMBA ck on the G logically in originally li e easier to g with the s quirements, etically. The system over hen an age hods and re to populate rio so users eded for the ject No. 0 G UI, articula stead of sted interpret if i hortest time current administrat time. ncy was sources for e each new only need t current 8-94 ting t or is ach o

Final Re RTC staf requirem useful to new meth identifies presents recomme FIGURE 30 Virginia VDOT te comment system. V provided definition populated VDOT co for bicyc listed as capital in agreed w changed VDOT al tool has a method n port f also identi ents and per evaluate the ods. This id the addition a snapshot o ndations. : RTC CURR Departme sted four sc s on the GU DOT repor were helpfu s. Since VD the remain mmented o le and pedes a safety and vestments m ith VDOT’s the tool. so observed feature to d ame in the m fied that the formance m current me entifies the al benefits f the RTC c ENT METHOD nt of Tran enarios and I and the tec ted that the l, but that n OT’s testin ing data item n the tool’s trian capital health progr ay occasion assessment some issue isplay more ethod brea current met etrics identi thods and re sufficiency obtained fro urrent metho S AND RESO sportation provided sev hnical cont layout and d ot all data it g, the resear s with defi categorizati investment am rather th ally fall und that it shou s with navig details abo kdown table 57 hods and re fied for a pr sources usin of the curren m implemen ds and reso URCES SCOR eral helpfu ent of the efinitions ems had ch team has nitions in th ons. Two of s. VDOT sta an a transpo er both pro ld be identif ation. The r ut each meth . This featur sources may ogram. As a g the same t methods t ting the new urces scored ING EXAMPL l e tool. the scenario ff questione rtation prog gram catego ied as a tran ecommenda od (Figure e presents i Pro address som comparativ scoring cons o address th method. F for compa E s that VDO d why this ram. Bicyc ries, but the sportation p tions page o 31) by click nformation “Generally, of tool is ject No. 0 e or all the e measure, i idered for t e program a igure 30 rison to the T tested we program wa le and pedes research tea rogram and f the softwa ing on the on each met the design great.” 8-94 t is he nd re s trian m re hod;

Final Re details th reference were not tool’s ref reviewed FIGURE 31 VDOT st than a sin volumes The weig rather tha agrees w added ca of the sof VDOT st users hav be useful users on current m for this ty port e relationsh s for additio sure how to erence guid at any time : TFGUIDE R aff also reco gle weight were selecte hts on these n setting bik ith VDOT th pability, so t tware tool. aff also note e adopted o . However, which recom ethods and pe of user f ips of each m nal informa access this e, with detai . EFERENCE G mmended a on the categ d as perform measures is e miles hig at this may his change w d that it wo r are current obtaining th mendations resources us eedback, bu ethod to re tion. VDOT information ls about all UIDE—EXAM llowing wei ory of measu ance metric set for all m her or lower be useful lo as not mad uld be usefu ly using. Th ese data wou they have i ers identify t it could be 58 quirements, staff noted in the tool. features in T PLE PARTIAL ghts on indi res. For ex s under the easures in than nonmo ng term, but e; however, l to identify e research t ld require c mplemented in the syste adapted in t metrics, and that this inf The research FGuide, so METHOD DE vidual perfo ample, bike health and s the health an torized volu the level of it can be co which reco eam agrees t ollecting fo or obtainin m. The curre he future to Pro resources; ormation is team inclu that method SCRIPTION rmance mea miles and n afety measu d safety me mes. The r effort is hig nsidered for mmended m hat this info llow-up info g permissio nt system i provide thi ject No. 0 and present useful, but s ded the soft s can be sures, rathe onmotorized res category asures categ esearch team h relative to future vers ethods othe rmation wo rmation fro n to share th s not design s informatio 8-94 s taff ware r . ory the ions r uld m e ed n.

Final Report Project No. 08-94 59 Finally, VDOT staff also requested an option to save scenario results to a PDF. The current system allows printing to any valid printer, including a PDF option. The user guide highlights how to export content to PDF. Dunbar Transportation Consulting DTC tested two scenarios: one for a large metropolitan agency and one for a statewide agency. DTC thought the selection options were straightforward, but indicated that some options may be too detailed for some users. The research team considered this feedback and thinks that the parallel User Guide can provide additional background for less experienced users and that this will help to bridge any knowledge gaps. DTC also thought that the “deselect” option on the performance metrics page was confusing. Most performance metrics have the option to remove a selection by clicking on the selected item, but a few were identified with a separate “deselect choice” button. To avoid confusion, these deselect buttons have been removed in favor of deselecting an option by clicking on the selection to remove it from the selection. DTC also observed that some recommendations were outside the stated budget, but with higher scores than those recommendations that were within the stated budget. This was an intentional part of the overall design for TFGuide; it does not currently consider any relationship between cost and budget. The research team has since revised how the cost affects the overall score by calculating a ratio of budget to cost; rather than just scoring recommendations within the budget higher than those outside the budget, this ratio is used in the score. The ratio concept helps to identify recommendations that are closer to the stated budget versus recommendations that are well beyond the stated budget. In addition, DTC identified two elements on the recommendations page that did not have pop-up explanations: industry adoption and contribution. DTC also suggested giving more prominence to the weights on the recommendations page. Based on this feedback, the research team included weights on the requirements and performance metrics pages so that these are considered before any recommendations are produced; the weights can be adjusted when reviewing the recommendations. This will allow a greater consideration of weights and refined recommendations as part of the requirements, rather than after the initial recommendations are produced. DTC also commented on the need to provide background information and references for travel forecasting practitioners who wish to pursue one or more of the recommendations. The User Guide provides this additional information to educate practitioners, planners, and decision- makers on the recommendations. The guidance is a starting point and is not intended to be a tutorial on how to implement the recommendation. The information provided in the User Guide is intended to provide enough background on a method to evaluate its appropriateness for a plan or program.