A brief history of airline booking systems

Thanks, airlines. Image: Getty.

Are algorithms trying to steal our children? That’s the question at the heart of the recent airline ticket pricing scandal. Are budget airlines trying to squeeze more money out of families by deliberately allocating them seats in separate parts of the plane, then forcing them to pay a fee so that children can be seated next to their responsible adult?

The airlines claim that when the fee isn’t paid, seats are simply randomly allocated. When the BBC’s Watchdog programme did some trials, a statistician claimed the result – in which every single person who booked was assigned an undesirable middle seat – was less likely than a National Lottery win.

As far as price gouging goes, airlines certainly have the means, motive, and opportunity. Plane ticket booking systems are a science developed over decades, and before they could rely on computers there were all sorts of wild and wonderful systems and contraptions: simple paper ledgers were replaced with index cards, which became blackboards, which became vast systems of boards of coloured chips. There was even a mechanical device that represented flight capacity, with tubes representing flights, filled with tiny plastic balls representing the seats. When a seat was booked on a fight, a button was pressed and a ball would pop out of the corresponding tube.

Algorithmic price adjustments have become a key part of this model, enabled by the vast amount of data that an airline booking system has to track just to function. Airlines have a huge incentive to maximise the number of people they can book on to any given flight and the amount of money they can extract from any given individual, too. While a passenger may be presented with a simple choice between economy and business, an airline has a far more sophisticated model based on how many tickets it can sell at any given price – which is why prices can fluctuate so wildly on online booking sites.

The person in the seat next to you may to all intents and purposes have the same flying experience, but as far as the airline is concerned will have bought a completely different ‘product’. As well as how far ahead a ticket is purchased, it’ll also factor in the likelihood that a ticket is being bought for business or pleasure: if it believes a ticket bought last minute midweek is most likely to be bought by someone on a work trip, it’ll jack up the prices.


This has inevitably led to an attitude of deep suspicion towards these companies. The jury is out on whether the widely believed idea that companies use browser cookies to hike the price on a second visit is actually something that really happens: one study found that, in 59 per cent of cases, checking the price again anonymously actually resulted in a higher price than if not trying to mask the user’s identity. In fact, it could just be a consequence of how volatile every other part of the pricing system is: but everyone still believes it anyway.

Whether or not this latest controversy is deliberate attempt to gouge people or the result of an algorithm that’s not been properly thought through, in an industry that’s grown up around gaming the system it wouldn’t be particularly shocking if someone had tried to do something like this to coin in a few more booking fees. On the face of it doesn’t seem like all that a great strategy: aside from the existing PR backlash from anxious families, it doesn’t take much imagination to ponder the, at best, tiresome, scenarios that might result from randomly seating unaccompanied five-year-olds on a packed flight.

The UK Civil Aviation Authority policy on this is that children should never be seated more than one seat away from a responsible adult: some of the budget airlines have simply made paying a fee a requirement for families. Presumably the algorithms will report the cost-benefit analysis of this grimy behaviour and adjust the fees accordingly.

There might be other, more profitable avenues anyway. Perhaps you could accept a cash payment from unruly teenagers to be moved away from the square parents they’re sick of after a two week holiday? Or charge people travelling on business a small fee to be surreptitiously parted from a particularly flatulent colleague? And what do the algorithms have to say about sedating passengers and stacking them up in coffins?

This article previously appeared on our sister site, the New Statesman.

 
 
 
 

The best bike maps are made by volunteers

A cyclist in Vancouver, Canada. Image: Getty.

Not all bike routes are equal. Some places that are marked as bike routes on a map feel precarious when traversed on two wheels, including shoulders covered in debris and places where you can feel the wind from speeding cars.

North American cities are building more bicycling routes, by adding on-street painted lanes, physically separated cycle tracks, bicycle-only or multi-use paths and local street bikeways. These different kinds of routes appeal to different types of users, from the interested but concerned cyclist to the keen road rider.

Despite this boost in biking infrastructure, a city’s website may not immediately reflect the changes or it may lack important information that can make cycling safer or more enjoyable.

Web-based maps that allow people to add information about bike routes give riders detailed data about the type of route, what it might feel like to ride there (do you have to ride close to cars?) and where it can take them (for example, shopping, work or school).

They can also tell us which cities are the most bike-friendly.

Measuring bike routes

We set out to assemble a dataset of bike routes in Canadian cities using their open data websites. But we found it was nearly impossible to keep it up-to-date because cities are constantly changing and the data are shared using different standards.

A physically separated cycle track in Victoria, British Columbia. Image: E. Gatti (TeamInteract.ca).

The solution was OpenStreetMap, which creates and distributes free geographic data. Anyone can add data or make edits to OpenStreetMap, whether they want to build a better bike map or make a navigation app.

We looked at OpenStreetMap data for three large cities (Vancouver, Toronto and Montréal) and three mid-sized cities (Victoria, Kelowna and Halifax) in Canada.

Not only did the data in OpenStreetMap agree reasonably well with the cities’ open data: in many cases it was more up-to-date. OpenStreetMap tended to include more local details such as where painted bike lanes ended and often marked the short cuts connecting suburban streets.

How did OpenStreetMap measure up?

Our analysis focused on how well different types of routes were mapped. We measured cycle tracks (which physically separate bikes from motorised traffic), on-street painted bike lanes (which use painted lines to separate bikes from motorised traffic), bike paths (which are located away from streets) and local street bikeways (which include traffic-calming features and where bicycling is encouraged).

Painted bike lanes are the most common type of route and also the most consistently well mapped. This makes sense, because the definition of a painted bike lane may be clearest across time and place. There is also a straightforward way for volunteers to tag it on OpenStreetMap.

But it was harder for us to distinguish cycle tracks from on-street painted lanes or paths (bicycle-only or multi-use) using OpenStreetMap. Local street bikeways were challenging to identify because of the wide range of ways cities design these kinds of routes along residential roads. Some use traffic-calming measures such as curb extensions, traffic islands, speed humps and raised traffic crossings to slow vehicle traffic and encourage safety, or greenery, reduced speed limits and bike-friendly markings on signs and the road surface.

Correspondence between OpenStreetMap and Open Data for categories of bicycling infrastructure. Image: author provided.

Bicycle routes that are physically separated from motor vehicles and pedestrians, like cycle tracks and bicycle-only paths, have the greatest benefits for bicycling safety and encourage bike use.

Ease of access to bicycle routes is important to a city’s overall bicycle friendliness, but there are other important things to consider including the distance to destinations, the number, slope and length of hills, number of riders and how the transportation culture of a city can influence its safety.


Bike-friendly Canadian cities

Our results showed that Montréal has the greatest total distance in cycle tracks in Canada. As cities continue building more bicycle routes, researchers and planners can use OpenStreetMap to measure these changes on the ground.

The perfect bicycle map is up-to-date, covers the entire globe and gives riders an idea of the kinds of experiences to expect on different trails, roads and paths. People cycling in cities can contribute to the high-quality geographic data needed to understand changes in bicycle friendliness.

But OpenStreetMap is only as good as its contributions. The exciting thing is that anyone who wants a better bike map — city planners, researchers and everyday riders — can join the bike-mapping revolution by logging in to OpenStreetMap and mapping the features that are important to bicyclists.

The Conversation

Colin Ferster, Post-doctoral fellow, University of Victoria and Meghan Winters, Associate Professor, Faculty of Health Sciences, Simon Fraser University

This article is republished from The Conversation under a Creative Commons license. Read the original article.