Welcome to Innaxis News

The Innaxis Post

Read our Blog

On maps

How are “mobility” and trips visualized and represented? Well, the most direct, intuitive way of doing so, is using maps. Representations, converting the 3-dimensional earth (*sphere*) to a flat  2-dimensional surface. This post is about maps, map properties, map distortion and curious maps. We hope you enjoy it!

Mapping the earth, or parts of it, is a classic, well-studied problem. For hundreds / thousands of years, cartographers and mathematicians have come up with different methods to map the curved surface of the earth to a flat plane. The main problem is that you cannot do this perfectly, (Theorema Egregium). The shape, area, distances and directions of the surface cannot be represented properly at the same time on a map.

1

Shape: If a map preserves shape, then feature outlines (like country boundaries or the coast lines) look the same on the map as they do on the earth. A map that preserves shape is conformal. The amount of distortion, however, is regular along some lines in the map. For example, features lying on the 20th parallel are equally distorted, features on the 40th parallel are equally distorted (but differently from those on the 20th parallel), and so on. The Mercator projection is one of the most famous and well-used shape-preserving maps:

2

Area: If a map preserves area, then the size of a feature on a map is the same relative to its size on the earth. For example, on an equal-area world map, Spain takes up the same percentage of map space that the actual Spain takes up on the earth. In an equal-area map, the shapes of most features are distorted. No map can preserve both shape and area for the whole world, although some come close over sizeable regions. Sinusoidal projection is an area-preserving projection:

3

Distance: If a line from a to b on a map is the same distance (accounting for scale) that it is on the earth, then the map line has true scale. No map has true scale everywhere, but most maps have at least one or two lines of true scale. For instance, in the Casini projection, the distances perpendicular to central meridian are preserved:

4

Direction: Direction, or azimuth, is measured in degrees of angle from north. On the earth, this means that the direction from a to b is the angle between the meridian on which a lies and the great circle arc connecting a to b. If the azimuth value from a to b is the same on a map as on the earth, then the map preserves direction from a to b. No map has true direction everywhere.

Finding the compromise: Most of the maps used are compromise solutions, partially preserving some of the above mentioned properties. The most used one (by far) is the one you can find in Google Maps, OpenStreetMaps etc. called Web Mercator, Google Web Mercator, WGS 84 Web Mercator or WGS 84/Pseudo-Mercator. It is a variation of the Mercator projection, ignoring the ellipticity of Earth for faster computation:

5

The Winkel tripel projection. “Triple” stands for trying to minimize errors in three properties at the same time: area, direction, and distance. The Winkel tripel is the arithmetic mean of different projections (equi-rectangular, area and shape preserving)

6

There is even a whole family called Myriahedral projections. These consider the earth *sphere* to be a polyhedron with a very large number of faces, a “myriahedron”. This myriahedron is cut open into small pieces and unfolded. The resulting maps have a large number of subareas that are (almost) conformal and that (almost) conserve areas. The location of the map interruptions can be “selected” (oftenly using sea areas etc)

7 8

 

 

 

 

 

 

 

 

Some ingenious representations mix the approach from Myriahedral projections and other property-preserving projections. e.g. the Goode Homolosineprojection:

9

All the previous projections provoke distorsion. There is no perfect projection. In the nineteenth century, Nicolas Auguste Tissot developed a simple method for analysing map-projection distortion. An infinitely small circle on the earth’s surface will be projected as an infinitely small ellipse on any given map projection. The resulting ellipse of distortion, or “indicatrix”, shows the amount and type of distortion at the location of the ellipse. Some examples for the most-used projections are given below.

10

If all the previous was not enough, it just leaves the door open to other projections that represent additional variables in maps, such as socio-demographic or technical indicators.

A map with country size proportional to population:

11

Proportional to number of immigrants:

12

Proportional to the number of tourists (Spain the biggest country in the world!):

13

Or even proportional to the total number of flights (this is one of my favourites!):

14

Some references and further reading on the topic:

How do you catch a plane?

How do you catch a plane? More interestingly: what are the stages of your door-to-door journey when an airline flight is involved? They’re most likely not all the same as anyone else’s.

There are a myriad ways of getting from your starting point (home, office, hotel) to the airport – from “door” to “kerb”. It could be by private car – whether “kiss-and-fly” (driven by family or friend), a ride-share with another passenger/co-worker, a taxi/minicab or their modern app-based equivalents, or we can just drive ourselves and park in the long-term or short-term car park (or maybe an off-site car park which then took you to your terminal by shuttle bus). It could be by bicycle or motorbike. It could be by bus, coach, tram, train, metro, or a combination of these – but how did you get to the stop/station for first one; walking, taxi, driving, “kiss-and-ride”, cycling – and where did the last one drop you off: the terminal, the airport transport hub? You may have driven, and then had to return, a hire car (if this is a return leg or your trip) and then took the hire company’s shuttle. If you left from an airport hotel, you most probably took their shuttle.

2-1

Getting through the airport (from “kerb” to “gate”) also involves many options. Did you check in online or do you have to do it at the terminal? Do you have bags to check in? Do you have to go through passport control? How long is the queue at the security check? How much buffer time did you leave, that you can now spend browsing around the shops? How many miles do you have to walk to get to your gate? Is your flight delayed?

Your answers to these questions (and their equivalents for the “gate” to “kerb” and “kerb” to “door” legs once your flight has landed), as well as the process for the actual flight(s) from gate-to-gate (including any transfers), have a bearing on how long your total journey will take. To be able to determine where there is room for reduction in this journey-time, and where research must be directed to initiate this reduction, in order to meet the ACARE goal of 4 hours door-to-door for intra-EU journeys, the DATASET 2050 team have analysed the component times of the current air journey in detail. This work is presented in the project’s deliverable 4.1 – Current Supply Profile.

Data for such analysis is hard to come by. Much of it is proprietary and, if it’s available at all, is sold at a high price – too high for this project. That which is available generally concerns all passengers, rather than just those on intra-EU travel; are people really likely to ride on one of the scheduled overnight coaches from Edinburgh to Heathrow to take a short-haul flight?

Making use of tools such as those provided by Google Maps, DATASET2050 researchers have been able to see the time taken to access airports by car, bicycle and public transport for a selection of airports.

2-2 2-3 2-4
 Berlin Tegel access-egress times by (L-R) bicycle, public transport, and car

(NB: scales are different, see them full size below)

These results and the many others included in D4.1 will help colleagues working on the next steps in the DATASET2050 project determine which parts of the different segments of your door-to-door journey can be speeded up, and where research and development is needed to further reduce our journey times.

 

 

2-2

2-3

2-4

Airport Economic Value study published!

The Modelling Airport Economic Value Study recently published (link here) has been made by the University of Westminster (Andrew Cook, Gerald Gurtner, Graham Tanner and Anne Graham) and Innaxis Research Institute (Samuel Cristobal), supporting EUROCONTROL (Denis Huet and Bruno Desart) within SESAR Project 06.03.01. The study provides a better understanding of the interdependencies of various key performance indicators (KPIs) and assesses the existence and behaviour of an airport economic optimum, in a similar way to the early 2000s, when estimating the economic en-route capacity optimum.

post1

By gathering for the first time real operational, financial and passenger-satisfaction-related data over 32 European airports, it was possible to develop and calibrate a model which produces reliable and realistic results. The fully calibrated results show the presence of a trade-off between the cost of extra capacity and the increase in the number of flights operated. As a consequence, all 32 airports exhibit a maximum in net income as a function of capacity, when the marginal cost of operating extra capacity is sufficiently low. This threshold in the marginal cost is, however, rather different across airports, and only a few airports can sustain a high cost of capacity: these are the largest and most congested airports, which clearly need extra capacity. This threshold is roughly consistent with the airports’ current operational cost of capacity, which means that they should be able to manage this growth, subject to the availability of investment.

The team has also developed a tool that provides access to all the features of the mathematical model with out having to dig into the equations. The underlying mathematical module is written in Python, while the interface is written in Matlab. The communication between the modules is transparent to the user and the software is capable of auto-calibrate the airport model using the existing or new data. The tool is flexible to explore the parameter space and different views of the output variables can be selected for a better understanding of the model outcomes. Results can be saved then in common format for further use (txt, csv, png, fig, etc.)

Screen Shot 2017-02-01 at 13.59.56 sample2122 blog_post32143

 

We warmly invite you to read the full report here!

Congrats to Samuel/UoW colleagues for such a superb study 😉

 

A new generation of business traveller

The DATASET2050 project does not only examine current European passenger profiles but also looks at possible passenger types in 2035 and 2050. To develop future demand profiles, current ones are either adjusted (see Current European PAX profiles), or new profiles are developed. As there is still a lot of uncertainty regarding how we are going to live and travel in the future, and since the project follows a data-driven approach, only passenger characteristics that can be supported by data are taken into consideration. Examples for developments supported by data are the ageing population in Europe, the increase in single households and the tendency to have fewer children per household.

For 2035, six future passenger profiles for the EU28 and EFTA countries are developed. Among these, the Digital Native Business Traveller was identified as one of the main passenger types in Europe. This group takes a journey mainly for occupational reasons and it can be seen as the new generation of business traveller. However, due to the high usage of technological devices one can assume that this passenger type is constantly connected and always online in continuous digital exchange with the private life, friends and family. He or she will be in the typically age of the working population of around 24 to 64 years, which today represents the digital savvy Generation Y and Generation Z. The income level and amount for transport expenditure will be medium to high. 0.5 to 1.5 trips per capita per year are taken, either alone or accompanied by another person. A large share of this passenger type will be female as the increase in female tertiary education enrolments might lead to an increase in working women within higher professions and hence an increase in women travelling for business purposes. Finally, he or she does not mind checking in luggage but takes only hand luggage when going on short trips. Public transport, taxi or car sharing are the preferred airport access mode choices.

3-1

Figure: The new generation of business traveller is digital savvy and constantly connected, enabled by emerging technology and new innovations to come.

This is one example of how a typical passenger group in 2035 could possibly look like. The outcome of all passenger demand profiles will be put in contrast with coming work packages (i.e. future supply profile), enabling this way a complete assessment on the European door-to-door mobility in the future. More information about the remaining passenger types, the methodology and databases can be found in the report on future passenger profiles.

INXmas greetings, 2017

We have had lots of fun innovating in 2016, so we are eager for a 2017 full of harder technical and scientific challenges, new research threads and complex innovation.

All the Innaxis team wish you a Merry Xmas break -including some fun and rest-  and a superb 2017!

ho ho ho!!!

img_7887

SafeClouds at EASA 2016 annual event and European Commission newsletter

SafeClouds.eu, the most advanced project to improve aviation safety through data analysis, was presented at the EASA 2016 Annual Safety Conference, held in Bratislava last November.
Carlos Alvarez, President of Innaxis, participated in the panel “Sharing and processing safety data: a vital step forward for safety?”. Carlos laid out the main goals of the project as well as our priorities for the next month, strengthening the importance of an integrated data pipeline, from low level raw data management to embedded analytics, driven by user operational questions. The integrated approach will be capable of developing data science solutions to provide all-new capabilities for safety improvements to aviation stakeholders.
You can watch the video of the session:
 shutterstock_245210677_plane_sunset_web_0
In parallel, SafeClouds.eu was also selected for the INEA/European Commission newsletter. This newsletter highlighted just 6 out of the hundreds projects recently awarded within the EU H2020 programme. As it is pointed in the newsletter “The (SafeClouds) project will develop a novel data mining approach for aviation safety and design innovative representations of the results in order to effectively transfer the gained to such users as airlines and air navigation service providers”

Current European PAX Profiles

Have you ever wondered about all the different people at the airport? Almost all of us have already flown: for going on holidays, visiting friends and family or going on a business trip. Likewise, many have been sitting at the airport, waiting at the gate and watching different passengers walking past. An airport is a melting pot where people of all ages, backgrounds, income levels and interests come together. As part of the DATASET2050 project, passenger characteristics are examined and six general passenger profiles (PAX profiles) are generated to gain an understanding of what distinguishes current European air travellers.

These PAX profiles are derived using existing passenger studies as well as data on demographical, geographical, socio-economic and behavioural aspects. At first, profiles are distinguished by travel purpose, i.e. whether passenger travel for personal or for business reasons. Since the amount of passengers travelling for private reasons exceeds that of passengers travelling for business reasons (on average across all EU28 + EFTA countries ten per cent business trips), there are four groups describing leisure passengers and two groups describing business travellers, as can be seen in the figure below. Following, passenger groups are assigned to pre-defined age intervals taken from an analysis of European countries as well as respective average travel activity within the particular age group.

4-1

Figure: PAX profiles according to travel purpose and age intervals with example profile information for “Executives”, “Family and Holiday Traveller” and “Best Agers” (own depiction based on PAX profile analysis)

All six passenger groups also differ by their income level. “Executives” have a high income; “Youngsters” have a low income and the remaining passenger groups have a medium income. Income alone has a great impact on travel budget and consequently on travel behaviour, i.e. how often someone is travelling or which transport mode is used to access the airport. Furthermore, the use of technical devices throughout the entire journey depends on age groups. Hence, all six passenger groups differ by the level of frequency in regard to mobile phone and internet usage. This translates to their booking and travelling behaviour as well. “Youngsters” and “Executives” are the two passenger groups using information and communication technologies (ICT) with a high frequency. “Youngsters”, for example, are digitally savvy and more likely to complete travel related tasks online compared to the group of “Best Agers”. Such processes along the journey could be online check-in or generating a boarding card on a mobile device.

The value of time also influences travel behaviour as passengers who value time a lot tend to save time along their journey and vice versa. Among all six PAX profiles, “Executives” and “Price-conscious Business Traveller” value time the most which is reflected, for instance, by their time-saving choice of hand luggage only. In contrast, “Youngsters” are young, often students or apprentices, and time rich but money poor. To compensate their low income, they tend to use public transport (often the longer access mode choice) to save money as they do not mind the additional time spent in public transport. “Family and Holiday Traveller” and “Best Agers” also have a rather low value of time.

The six passenger groups also differ by their length of stay. The trip length in terms of nights staying is another parameter influencing the amount of luggage a particular passenger is taking along the journey. The amount of nights spent at a particular destination differs both by travel purpose and by type of journey conducted. Business travellers tend to spend fewer nights per trip than leisure passengers. And “Youngsters” visiting friends in urban centres spend less nights than “Family and Holiday Traveller” on their summer vacation. In turn, this may influence the access mode selected, the time spent in luggage check-in processes, or during luggage collection at the destination airport. For instance, in order to minimize time and effort accrued to respective handling processes, business passengers reduce the amount of luggage taken along. Finally, it is important to mention that one person can be assigned to several PAX profiles. A manager of an international company can travel for business purposes (being assigned to the group of “Executives”) and in private life being a dad and flying with his wife and two children into the summer vacation (being a “Family and Holiday Traveller”).

More information on the PAX profiles and the analysis can be found in the DATASET2050 report “Data driven approach for a Seamless Efficient Travelling in 2050”.

ComplexityCosts holds Close-Out Meeting

After more than four years of intensive research the SESAR WP-E project ComplexityCosts has finally concluded. Project partners from the University of Westminster and Innaxis gathered with experts at the EUROCONTROL Experimental Centre in Brétigny-sur-Orge, France in early October to discuss the project results and close the research activities.

The ComplexityCosts study focused on the trade-offs between several disturbances and mechanisms from a passenger and stakeholder’s costs perspective. A major highlight of the project was the concept of resilience costs and further identifying a metric to effectively measure it.

The research continues on within the Horizon2020 DATASET2050 and SESAR LTER Vista projects, so stay tuned for further developments on mobility research.

SafeClouds kick off meeting

safeclouds-kom

SafeClouds.eu, a H2020 big data for safety project, coordinated by Innaxis, kicked off earlier this month.

SafeClouds is the recently launched H2020 aviation-safety project. It is coordinated by Innaxis, with 15 additional entities (including airlines, ANSPs, EASA, Eurocontrol, various research entities, etc) from 8 different countries.
The aim of SafeClouds is to improve aviation safety by developing state-of-art big data and data analysis tools. The consortium will build a coordinated platform to combine and share data among different aviation actors.
img_6330

DATASET2050 presentation at Data Science in Aviation Workshop (EASA, Cologne-Germany)

The annual event exploring Data Science in Aviation (ComplexWorld funded; organized by Innaxis) has recently celebrated its fourth edition this past September 8th 2016. The event was hosted on the EASA premises in Cologne, Germany . This year it highlighted a presentation of the DATASET2050 project, “Data Science for Mobility”, by project coordinator Samuel Cristobal (Innaxis).

5-1

 

 

On the Data Science In Aviation event:

Previous editions of the Data Science in Aviation event were hosted in Madrid 2013, Paris 2014 and Brussels 2015. The popular event usually draws attendance from more than 80 individuals from top European and worldwide aviation entities (including Airbus, Eurocontrol, Boeing, EASA, Airlines, Airports, ANSPs, SESAR , Universities, etc) along with ICT and data-related entities (including CERN researchers, Fraunhofer, Infrastructure-related, and various universities). Notable presenters from the 2016 edition included EASA, Innaxis, NATS, Eurocontrol, Boeing, ENAC and Fraunhofer.

In terms of the event agenda and content, the presentations has traditionally outlined how data science is understood as a useful set of fundamental principles that support and guide the principled extraction of information and knowledge from aviation data. Furthermore, the discipline leans on well-known data-mining techniques, and goes far beyond these techniques with successful data-science paradigms which provide specific applications in various air transport areas (safety, performance, mobility etc).

 

On DATASET2050 Samuel’s presentation:

The event also highlighted a key presentation from Innaxis project coordinator, Samuel Cristobal. Samuel presented five different points on data science in aviation.

  • First, he explained how some of the data science tools, techniques and concepts have been used in the mobility context, specifically using the DATASET2050 project as a case example.
  • Second, Samuel explained the different door to door phases under analysis (door-kerb; kerb-gate; gate-gate; kerb-door), which helps to delve deeper in the different data science components within aviation phases.
  • Third, Samuel outlined the different links between project objectives and overall Flightpath2050 goals.
  • The fourth point explored mobility data in Europe, and the value of the DATASET approach in this context.
  • The presentation concluded with a fifth and final point announcing the next communication actions. The full presentation can be accessed here: https://www.dropbox.com/s/91julyl8gsij2k9/DATASET_SC_v1.pdf?dl=0

 

In sum, the fourth edition of the Data Science in Aviation event was an excellent opportunity for dissemination of DATASET2050. This was in conjunction with a fruitful exchange of ideas with other aviation data scientists, some of whom working with similar tools in other sub-areas far from mobility. We hope to continue this momentum of knowledge exchange and look forward to a potential fifth edition of the popular event.

 

You can watch DATASET2050 Samuel’s presentation here, and the rest of the event videos at Innaxis’ Vimeo channel.

Connect with us!