Presentation on theme: "ETISplus Air transport Data collection & selected results Oliver Schnell MKmetric GmbH 05.11.2012- Brussels."— Presentation transcript:
ETISplus Air transport Data collection & selected results Oliver Schnell MKmetric GmbH Brussels
2 Fields of Data Collection in Air Transport Air network (WP7), consisting of - Airports, including supply indicators - Air links between airports generated - Impedances calculated => In principal data validation already takes place within the process of generating the air network => But internal ETIS+ evaluation showed some need for improving check algorithms for flight schedule input (times, daylight saving times, stopovers) Air passenger (WP9) and air cargo (WP8) flows - link loads - node loads => In principal data validation already takes place within the process of generating the air passenger and cargo flows -
Approach to built the air networks (2005&2010) Harmonised Statistics for links (airport pairs) in tons / passengers from EUROSTAT & other sources Flights from airlines, airports, tour operators and commercial schedule data provider Flight database to calculate paths airport airport MKmetric Connection builder Airport-airport-connections Additional (charter) flights not covered by OAG mn aggregation tool (MKm) Regionalisation for ETISplus Europe: NUTS3, US etc: Federal States, Rest of World: countries / states (NEA) Airports to be considered as O & D Carrier & airport specific minimum connecting / transhipment times to collect Air network (i.e. Airport-airport-legs including path information (sequence of nodes) Region-Region impedances (time, distance,...) KIT (Mixed) links region airport APC Airport / route choice tool (MKm) as part of the calibration process (WP9) for costs Dark green: external input Light green: adjusted external input Light blue: intermedia te output to be further used in the approach Dark blue: Ouput Orange: MKmetric internal tools As schedule and statistics are published regularly: repeatable process feasible year by year!
4 Checks done when generating the network Flight schedules may vary day by day, but calculating airport-airport connections for each day of a year is not feasible. Solution: use schedules of typical week (2 nd week in June). Resulting in: - enrich typical week by flights between airports where statistics exist but no schedules - transfer those flights offered only seasonally or ceased / introduced during the year into typical week with representative frequency (7/7 5 months => 3/7 avg.) Internal validation showed: - for some flights incorrect schedules were provided from commercial sources (flight times to long / to short in opposite direction (e.g. Manchester – Goa, India) - variants in country specific period when daylight saving times apply, may cause to short / to long flight times, when transferring flights into typical week - multi-stop flights may cause suspicious flight times (to long for given distance), so procedures for automatic corrections of flight times must deal with intermediate stops Solution: check flight schedules on air link speed, minimum / maximum times, arrival day info.
Air passenger: Approach scheme Harmonised passenger statistics for links and nodes (Eurostat, national statistics e.g. US, DE) Data base for calibration process Iterative calibration process to balance economic weights of regions, link loads, 3-point or destination statistics, node loads and impedances Regionalisation and corresponding socio-economic data (trip generation) Air passenger matrix between ETISplus regions for three trip purposes according to regional data Feeder links region airport (combined of rail & road feeders Additional statistics according to flights Air passenger network airport airport Air transport impedances region region WP7 Passengers surveys for trip purpose split and route choice On base of a start-up matrix (year before), building the matrix is a repeatable process
Iterative calibration process Distribution Generation Route Choice Assignment Link Load Assessment Matrix Evaluation Airport / Route Choice Assessment Basic Statistic Evaluation Matrix Bounding Model / Input Adjustment Statistic Evaluation Continuous validation is the core part of the matrix generating process The iterative process results in a flow matrix, where the overall sum of deviations between assigned link / node loads and link / airport statistics is minimal (but not 0) The same principal approach has been done for air cargo
7 Resulting networks delivered, Terminals The air networks delivered consists of three components: airport file, for each airport consisting of DATA_GROUP Metadata on dataset level, since the DB-system only supports metadata on table level. Observed, estimated, modeled data SOURCE Metadata on dataset level, since the DB-system only supports metadata on table level. Data provider, e.g. UN, EUROSTAT, NEA, IWW, MKm, TRT, … FROMDATE effective in time range; YEAR or period Fromdate, Todate, e.g. Fromdate , todate is preferred TODATE effective in time range; YEAR or period Fromdate, Todate, e.g. Fromdate , todate is preferred SEASON effective in time range; WINTER | SUMMER | ALL; so that this information need not be deduced by client programs NoteIDNumeric Identifier TLCOrigin [TLC) e.g. FRA Airport Name X_coordinate decimal longitude coordinate Y_coordinate decimal latitude coordinate Numerous indicators Supply side finished, demand finished, all delivered
Demand indicators air transport Outcome of the projects TOOLQIT, ETIS Base and WORLDNET has been scanned for indicators to be derived directly or indirectly from the networks. Further input came from the modeller workshop, Brussels, The following terminal indicators for the demand side of air transport have been derived based on the passenger / cargo matrices: -total number of passengers -total number of domestic passengers -total number of intercontinental passengers -total number of Schengen passengers -total number of European passengers -total number of North American passengers -total number of South and Middle American passengers -total number of Asian, Australian / Oceania passengers -total number of African passengers -total number of Middle East passengers -total number of air cargo tons
International Air Passengers 2010
10 Supply indicators air transport The following terminal indicators for the supply side of air transport were derived: -total number of aircraft movements passenger transport (PAS_CAF of EUROSTAT) -total number of air cargo liner movements (TOT_CAF \ PASS_CAF of EUROSTAT) -largest airplane (MTOW) which landed/started (MKm network) -total seat capacity offered (PAS_ST of EUROSTAT) -total cargo capacity offered (MKm network) -number of destinations offered (MKm network) -number of domestic destinations (MKm network) -number of intercontinental destinations (MKm network) -number of European destinations (MKm network) -number of Schengen destinations (MKm network) -airlines market position, by seats for top ten & other (MKm network) -alliance market position by seats Oneworld, Sky team, Star Alliance, Low Cost, other (MKm network)
Example: Airports by Carrier Type Seats offered by Alliance / carrier type gives information e.g. on competition issues
12 Resulting network delivered, Air Links Air links consisting of - Notes:- links include surface connections (e.g. LHR LGW) - Air connections like Frankfurt – Amsterdam – Manila may have different impedances concerning time frequency, price, etc. than the corresponding values added for Frankfurt – Amsterdam and Amsterdam – Manila! For that reason for each pair of airports the network consists of the direct link (if existing) or otherwise of the leg connecting it within shortest travel / transport time when requiring in maximum two stopover airports. IndicatorVerbal description of indicator DATA_GROUP Metadata on dataset level, since the DB-system only supports metadata on table level. Observed, estimated, modeled data SOURCE Metadata on dataset level, since the DB-system only supports metadata on table level. Observed, modeled data FROMDATE effective in time range; YEAR or period Fromdate, Todate, e.g. Fromdate , todate is preferred TODATE effective in time range; YEAR or period Fromdate, Todate, e.g. Fromdate , todate is preferred SEASON effective in time range; WINTER | SUMMER | ALL; so that this information need not be deduced by client programs ObjectIDIdentifier used by ArcGIS ShapeArcGIS geometry IDA unique identifier for each link FromNodelDFromnode for the link ToNodelDTonode for the link OpenFor Indicates whether the air link is open or not in the forward direction (fromnode to tonode) (0 or 1) OpenBack Indicates whether the air link is open or not in the backward direction (tonode to fromnode) (0 or 1) FreeSpeed1, not applicable QueueSpeed1, not applicable LanesFor1, not applicable LanesBack1, not applicable LinkTypelDType [cargo/pax/both] LaneHCFor-1, not applicable LaneHCBack-1, not applicable Active Indicates whether the link is included in the calculations or not (1=active, 0=inactive) Lengthdistance [km] OrigIATA-Code for origin airport. Not used in calculations Orig_NameOrigin airport name. Not used in calculations DestIATA-Code for destination airport. Not used in calculations Dest_NameDestination airport name. Not used in calculations CostBTicket fare for business trips (Euro) CostPTicket fare for private trips (Euro) CostHTicket fare for holiday trips (Euro) LinkTimeFlight time on link (Minutes) TransferTimeExtra transfer time on link (Minutes) TotalTravelTimeTotal_time [minutes] Annual_DepartNumber of flights per year (Departures per year) Daily_DepartNumber of flights per day (Departures per year) HeadwayTime Average waiting time due to frequency. Calculated based on Daily_Depart as: HeadwayTime = >2*(20/Daily_Depart)*60. Minimum 60 minutes and maximum 240 minutes (minutes) LinkClass Network type (FLC=Flight Connection, ETR=External Transfer between nearby airports). Not used in calculations LowCost Share of low cost connections [0 -> 100] Shape_Length ArcGIS link length. Unit determined by data projection. Default is degrees Transfer1node for the linkTransfer1node for the link [0 if n.a.] Transfer2node for the linkTransfer2node for the link [0 if n.a.) Transfer1Transfer 1 [TLC) "---" id n.a. Transfer2Transfer 2 [TLC) "---" if n.a. FrequencyPaxfrequency passenger [connections/week] FrequencyCargofrequency cargo [connections/week]
13 Resulting impedances delivered Concerning air transport, the impedances consist of the average values from up to 16 different routings considered in the APC routine and for which choice probabilities are calculated. For that reason we concentrate on - travel time (total and in vehicle) - cost (not for air cargo, as WORLDNET approach doesnt cope with) - frequency - distance Impedances have been used to calculate air accessibility. Indicator Description of indicator DATA_GROUPMetadata on dataset level: type of data SOURCEMetadata on dataset level: data provider FROMDATEOperating time. Year or starting date TODATEOperating time. Year or ending date SEASONSeason of operating time IDUnique identifier FromRegOrigin ETISZone3_ID ToRegDestination ETISZone3_ID PurposeTrip purpose (only for passenger) TotalTimeTotal travel time AccessTimeAccess time FlightTimeFlight time (including transfer times, check-in/-out etc.) AccessCostAccess costs (only for passenger) FlightCostFlight costs (only for passenger) FrequencyConnections per week AccessDistanceAccess distance FlightDistanceDistance of flight
Application: Air transport accessibility 2010 Displayed is the average travel-time for each NUTS3 region in Europe needed to reach another European region by air transport, including time for surface access / egress. Travel-times have been weighted by the population of each destination region; i.e. travel-time to reach a region in the North of Finland has a lower weight than the time to reach the city of London
MKmetric Gesellschaft für Systemplanung mbH Oliver Schnell Rüppurrer Straße Karlsruhe Germany Phone: Fax: webpage: Thank you for your attention