MEMORANDUM September 1, 2015 - FWS

2y ago
65 Views
2 Downloads
602.53 KB
5 Pages
Last View : 1m ago
Last Download : 2m ago
Upload by : Lilly Andre
Transcription

MEMORANDUMTOFROMS U BJ ECT S e p t e m b e r 1, 2 0 1 5Craig O ’Connor, N C A AlEcE6 - A ir Travel D ata IntegrationThis m em o describes how D B IB data from tire Bureau o f Transportation Statistics(D epartm ent o f Transportation) are used in the calculation o f travel costs for the tripsreported by respondents to the national and local valuation surveys. A s discussed inTechnical M emo E l-T ra v el Cost Com putation, one o f the com ponents o f travel cost isthe cost o f flying. The D B IB data are used to calculate the cost o f flying by providing anestim ate o f the m onetary cost and the tim e spent travelling (used in the opportunity costo f flying tim e calculation) for each potential route. A “route” is used here to represent aunique origin airport/destination airport pair. Each route has m ultiple itineraries in theDB IB dataset (described in detail below) that are sum m arized through this process toprovide one cost estim ate and one tim e estim ate per route. The direct cost o f a given routeis represented by the 30* percentile o f airfares for itineraries serving the specified route,plus baggage fees, w here applicable. The estim ated travel tim e for each route includesthe tim e spent in the airport (before and after the flights) and estim ates o f the flight tim eand layover duration fo r the itinerary associated w ith the 30* percentile airfare.DATASETSThe cost and tim e associated w ith each route were estim ated using data on available flightroutes, ticket prices and fees, and flight and layover durations. These elem ents weredraw n from the national airport and flight data sources described below. W here completeinform ation was not available for a relevant route, the data w ere im puted according to theprocedures detailed in the following section.AIRPORT ENPLANEMENTSThe FA A provides a list' o f all airports in the U nited States, along w ith their yearlyenplanem ents and hub size designation. Hub sizes are assigned based on percentage o fall annual passenger boardings in the US {Large— 1% or m ore; M edium — 0.25-1%;Sm all— 0.05-0.25% ; N on-hub— less than 0.05% but more than 10,000 enplanem ents).The hub size designations arc used in the layover duration estimate.A IR P O RT LOCATIONSA list o f airports w as extracted from the 2012 D O T Bureau o f Transportation Statistics(BTS) Airports database shapefile using A rcG lS. The resulting input dataset includes theairport code and location for all aircraft landing facilities in the U nited States and U.S. F ed eral A viation A dm inistration (FAA) C alen d ar Y ear 2012 P assen g er Boarding and All-Cargo D ata listsD W H-AR0056788

territories. Tliese location data w ere used to m atch respondent origins w ith the closestairports (resulting in the PC M iler dataset described below).AI RL INE O RI G I N AND D E ST I N A T I O N S UR VE Y ( D B 1 B )The A irline O rigin and D estination Survey is a 10% sample o f all airline tickets collectedby the Office o f A irline Inform ation and the Bureau o f Transportation Statistics for eachyear since 1993. The survey data are presented in three data tables {ticket, market, andcoupon) collectively referred to as DB IB. The three tables can be joined together by an“Itinerary ID” field. The ticket data table includes one observation per itinerary, be itone-w ay or round-trip, direct or indirect. It includes the origin airport, num ber o fpassengers, round-trip status, fares, carrier, and flight distance. It does not includeinform ation about the destination or layovers associated w ith the itinerary. Thatadditional inform ation is found in the m arket data table, in w hich each observationrepresents one direction o f a trip. Therefore, one-w ay tickets in the ticket table have onem atching record in the m arket data table, and round-trip tickets have tw o matchingrecords. The m arket table also lists the connection airports related to the trip, ifapplicable. One observation in the coupon data table represents one leg o f a trip,including separate observations for each leg o f a trip w ith layovers. This flight costanalysis draw s upon data from the ticket and m arket data tables for all airline ticketsstarting in the second quarter o f 2010, through the second quarter o f 2013.OAG F LI GH T TIMESFlight tim e data are provided by O A G A viation Solutions Schedules Database (reportedA pril 25, 2012). Each record in this dataset is an origin-destination-elapsed flig h t timetriplet for all direct flights w ithin the U.S. during 2011. The frequency o f each triplet isalso presented w hich allow s for w eighted statistics.S A B R E L A Y O V E R DATAThese data are from Sabre A irline Solutions. They include layover inform ation for a 500flight sample o f one-stop, indirect routes w ithin the contiguous U nited States thatoccurred during July o f 2011. It provides the passenger w eighted and un-w eightedm edian layover tim e by route, as w ell as the passenger w eighted m ean layover tim e foreach origin-connection-destination triplet.PROCEDURESThe final output o f the D B IB procedure is quarterly estim ated tim es and costs for eachtrip identified in the national and local valuation sur\ eys. These values were estim atedaccording to the follow ing procedures.PRELIMINARY DB 1B AND A IR P O RT LIST PROC ESS INGThe first step was to create a subset o f D B IB ticket and m arket data pertaining to roundtrip, direct flights. The m ain purpose o f this interim dataset w as to provide distances foreach direct flight for use in the O A G flight tim e estim ation equation. DB IB ticket andDBIB d a ta w e re dow nloaded from th e BTS w eb site on th e follow ing d a te s : Jan u ary 7, 2013 fo r th e 2010 and 2011 q u a rte rlyd a ta ; Jan u ary 6, 2013 for 2012 Q u a rte r 1; F eb ru ary 27, 2013 fo r 2012 Q u arters 2 and 3; April 16, 2013 fo r 2012 Q u a rte r 4;July 24, 2013 for 2013 Q u arter 1; an d O cto b e r 28, 2013 fo r 2013 Q u a rte r 2.D W H-AR0056789

m arket data tables were m erged, and itineraries representing direct, round-trip flightsw ithin the contiguous United States were kept. This ticket level dataset w as thencollapsed dow n to provide the passenger w eighted m edian distance and m iles-flow n foreach unique origin to destination route.In a separate process, the FAA enplanem ent and D O T A irport list were com bined andcleaned to produce an airport reference dataset w hich included airports in the contiguousU nited States w ith annual enplanem ents over 100,000 that were represented in bothsource datasets.OAG REGRESSIONBecause the OA G dataset did not include flight tim es for all relevant itineraries, aregression model was estim ated to define a relationship betw een flight distance andelapsed flight tim es th at was then used to im pute m issing tim es.From the original O A G dataset, observations were dropped from the analysis if theyrepresented routes outside o f the contiguous U nited States, or if they reported flight tim esover 600 m inutes. The dataset was then collapsed so that each route was represented bythe m edian flight tim e, w eighted by the frequency o f occurrence (num ber o f flights alongthe specified route lasting the specified time).The OA G analysis data were m erged w ith the direct flight and airport datasets to bring inadditional inform ation about the origin and destination airports, and flights (includingpassenger w eighted m edian distance).The relationship betw een distance and flight tim e was defined in a linear regression o fflight tim e on distance, where flight tim e is the m edian flight tim e from the OAG analysisdataset and distance is the passenger w eighted distance from the D B IB direct flights data.The estim ated intercept and distance param eter w ere used to im pute m issing flight tim esin later D B IB processing steps.EXHIBIT 1. OAG REGRESSION RESULTSY MEDIAN FLIGHT TIME (MINUTES)BP-VALUED i s t a n c e (ml)0.1213 0. 001Constant42.50 0. 001R-squared 0.9537S A B R E L A Y O V E R T I ME E S T I M A T E SThe connection hub airports listed in the Sabre dataset were m erged w ith the airport listto create a dataset that included the three Sabre route airports and additional inform ationon the connection airport, including hub size. This list o f unique routes was thensum m arized by hub size to calculate the passenger w eighted m edian o f the reportedpassenger w eighted m edian layover tim e per route for each hub size (small, medium, andlarge). These sum m ary statistics were incorporated into the calculation o f flight tim es asdescribed in the next step.DW H-AR0056790

EXHIBIT 2. SABRE LAYOVER TIMES BY HUB SIZEHUB SIZELAYOVER TIMELarge70Medium55S ma l l80NOTES: HUB SIZES ARE DESIGNATED BY THE FAA. LAYOVER TIME IS THE PASSENGER WEIGHTED MEDIAN LAYOVERTIME IN MINUTES.A S S IG N I N G A I RF A R E S AND T RAVE L TIMES TO R E P O R T E D T R I P R O U T E SFor each quarter and route, the 30* percentile airfare was determ ined using all relevantitineraries in the DB IB ticket and m arket datasets. To calculate flight tim e estim ates, theD B IB eligible itineraries w ere com bined w ith data obtained during OAG and Sabreanalyses. The resnlting flight tim e and layover dnration datasets with one observation peritinerary, separated by quarter, w ere then sum m arized to obtain estim ates for eachrelevant route in the national and local shoreline sam ples for each quarter.Ai r f a r e sFor a given route and quarter, relevant airfares w ere defined as those associated withitineraries that were round-trip; had distances, m iles-flow n and fares that were consistentbetw een ticket and m arket datasets and had airfares betw een 50 and 5000. A 50baggage fee was added to the airfare for all carriers except JetBlue and SouthwestAirlines.Travel tim esThe first elem ent o f travel time com puted was flight time. E ach itinerary was brokendow n by leg so that each leg origin and destination pair could be m atched w ith the flighttim es reported for direct flights in the OAG dataset, w hen available. W here the originand destination o f a leg m atched a route in the O A G dataset, the passenger w eightedm edian flight tim e w as recorded for that leg. W hen the leg was not in the OAG dataset,but was in the direct flights dataset, flight tim e w as estim ated using the O AG regressionresults:F lig h t ttvfieiggPdistance * DtstCLTLCeigg -j- PconstantFor itineraries w here every leg w ithin the route was estim ated using the above methods,air travel tim e w as aggregated across legs and assigned to the itinerary. If one or morelegs w ithin an itinerary were not in the O A G dataset or the direct flights dataset, the OAGregression param eters were used to estim ate total travel tim e by applying the distancecoefficient to the total distance traveled (inbound and outbound), and the interceptcoefficient to the num ber o f legs in the itinerary.F lig h t tijnenifigi-gify distance * J istcmcenijigj-gij-y constant *9 itineraryN ext, tim e spent during layovers w as added to the flight tim e, w here appropriate, basedon the connection size layover estim ates from the Sabre dataset. Connecting airportsw ere assigned a layover tim e based on the passenger w eighted m edian layover tim e (J,)DW H-AR0056791

associated w ith th eir designated hub size (S, M , or L). The total layover tim e per itinerarywas calcnlated as the num ber o f connections o f each hub size m ultiplied by the m edianw eighted layover tim e for the respective hub size:Luyovev timoifijigj-dj-yITi * Connectionsii-l,legsitineraryTw o hours w ere added to each route to account for pre- and post-flight tim e at the airport.E s tim a te d 30 th P e r c e n t i l e F ares and Travel T im es by R ou teThe fares datasets representing relevant itineraries by quarter and year w ere sum m arizedto the route level, creating an output o f one estim ated fare and travel tim e for each routein each quarter. Using the qnarterly fares datasets, the 30* percentile itinerary fare wascalculated for each origin-destination pair. I f this percentile exactly m atched a reportedfare for an itinerary covering th at route, that itineraiy' fare was selected to represent theroute fare. If none o f the itinerary fares m atched the 30* percentile, the route fare wascalculated as the passenger w eighted m ean o f itinerary fares ju st above and below the 30*percentile. A sim ilar process w as applied to estim ate travel tim es for each route. W herean exact m atch on itinerary fare to 30* percentile itinerary fare existed, the travel tim e o fthe m atched itinerary w as selected to represent that route (or the passenger w eightedm ean travel tim e o f all itineraries m atching the 30* percentile fare, if more than one exactfare m atched). W here none o f the itinerary fares m atched the 30* percentile fare exactly,the route tim e w as calculated as the passenger w eighted m ean o f the flight tim es o f theitineraries w ith fares ju s t above and below the 30* percentile.D W H-AR0056792

airports (resulting in the PC Miler dataset described below). AIRLINE ORIGIN AND DESTINATION SURVEY (DB1B) The Airline Origin and Destination Survey is a 10% sample of all airline tickets collected by the Office of Airline Information and the Bure

Related Documents:

Confidential Information Memorandum June 30, 2011 Sample Industries, Inc. (Not a real company.) Prepared by: John Smith, CPA Middle Market Business Advisors 500 North Michigan Ave. Chicago, IL. 60600 This Memorandum is confidential and private. Distribution is restricted.File Size: 211KBPage Count: 16Explore furtherInformation Memorandum Disclaimer - Free Template Sample .lawpath.com.auConfidential Information Memorandum (CIM): Detailed Guide .www.mergersandinquisitions.comInformation Memorandum Template for Investors Property .businessplans.com.auRecommended to you b

‘‘Memorandum’’ the memorandum of association of the Company, as amended from time to time ‘‘Memorandum Amendments’’ the proposed amendments to the Memorandum subject to the approval of the Shareholders by way of a special resolution at the EGM ‘‘Second Amended M&A’’ the second amended and restated Memorandum and Articles,

best practices for FWS implementation. Second, we summarize the literature and most significant studies examining the effects of participation in FWS. The aim of this effort is to understand the current research on FWS and the program's impact on students. Based on the literature, we identified the following overarching themes:

2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 . Removal handle Sound output / wax protection system. 11 Virto V-10 Custom made shell Battery door Volume control (optional) Push button Removal handle . Before using

Health Technical Memorandum 07-02: EnCO 2 de 2015 - making energy work in healthcare iv Structure of the Health Technical Memorandum suite The series contains a suite of nine core subjects: Health Technical Memorandum 00 Policies and principles (applicable to all Health Technical Memoranda in this series) Choice Framework for local Policy and

NATIONAL SENIOR CERTIFICATE GRADE 12 SEPTEMBER 2015 GEOGRAPHY P1 MEMORANDUM MARKS: 225 This memorandum consists of 15 pages.

3. "Floral Design Workshops" 4. 2006 Master Gardener Trainee Class Roundup 5. "Back to Your Roots," the Southeastern Regional Master Gardener Conference Happy September 1 September 14 September 15 September 17 September 19 September 20 September 24 September 26 September

Calendar of Torah and Haftarah Readings 5776-5778 - 2 - DATES OF FESTIVALS 5776 2015-16 5777 2016-17 5778 2017-18 Erev 1st Day Rosh Hashanah 1 Tishri 13 September 2 October 20 September 1st Day Rosh Hashanah 1 Tishri 14 September 3 October 21 September 2nd Day Rosh Hashanah 2 Tishri 15 September 4 October 22 September Kol Nidre 10 Tishri 22 September 11 October 29 September