杏吧原创

Crunch commuter data to track changing communities

Urban well-being could be boosted thanks to lessons learned from tracking people through a city's underground travel network
Where do you go to, my lovelies?
Where do you go to, my lovelies?
(Image: Lefteris Pitarakis/AP/PA)

Editorial:Train tracks of our tears put to good use

LONDON commuters are generally a surly bunch, grumbling as they battle through the city鈥檚 underground train network each morning. Nevertheless, records of their journeys could be a key to improving urban well-being.

Every day, millions of Londoners touch their Oyster card to the underground鈥檚 wireless ticket readers each time they enter and exit the system, building up a detailed database of travel through the city. Computer scientist Daniele Quercia and colleagues at the University of Cambridge have now compared this data with official measures of social deprivation and found that a community鈥檚 prosperity is reflected in the comings and goings of its residents.

Previous research has shown that when people are asked to describe their mental map of a city, the neighbourhoods they recall in better detail 鈥 areas with higher 鈥渧isibility鈥 鈥 tend to score higher in social and economic measures of well-being. Gathering such measures normally involves manual surveys, but the team wanted to know if people鈥檚 movements across the city could be used as an easily collected proxy.

They tested their idea by using one month of Oyster-card data 鈥 about 76 million underground and overground rail journeys. Oyster cards have anonymised ID numbers, so the team worked out roughly where commuters lived by picking the two most visited stations for each person, and assumed these corresponded to home and work locations.

Quercia and colleagues then compared the number of visitors an area received with the , a composite measure based on census data that weighs factors such as income, health and crime levels in a particular community. They found that more deprived areas were visited more often in general, but also discovered more subtle effects. People from deprived areas visited both other deprived areas and prosperous areas 鈥 defined as those with lower IMD scores. Residents of better-off communities, on the other hand, tended to only visit other privileged neighbourhoods. Not all Londoners use the underground, the team notes, so the lack of car users, say, may skew the results.

It may not be surprising that people in comfortable neighbourhoods keep to themselves. But Quercia says the results are still useful because they would allow policy-makers to respond more rapidly to changing social dynamics within a city than by relying on census data alone, which is taken just once a decade.

鈥淵ou have a real-time way to track changes,鈥 he says, adding that his technique could be used to quickly measure the effect of the London 2012 Olympics on surrounding neighbourhoods. 鈥淵ou couldn鈥檛 do that at all from the census data.鈥 The team will present its findings at the conference in Newcastle, UK, in June.

聯The technique could measure the effect of the 2012 Olympics on London鈥檚 neighbourhoods聰

Quercia鈥檚 team has also developed another way to collect real-time visibility data 鈥 an online game called that tests which places stick in Londoner鈥檚 minds. Players are shown a Google Street View image of a random location then asked to match it to the nearest tube station or borough. The extra data stream can be used in a similar way to the Oyster card touch-ins to provide insights to community well-being, Quercia claims. More data sources should lead to more accurate results. That could help city planners understand how certain events 鈥 a neighbourhood festival, say, or an extensive construction project 鈥 affect communities.

, a researcher at University College London who runs an iPhone app happiness survey called , says the work is an interesting idea but that deprivation scores do not inversely correlate precisely with well-being, though they are connected. Using transport data as a proxy may hold true on average, he says, but there is a danger in ascribing social meaning to every blip in that data. 鈥淚f you want to know people鈥檚 well-being, then asking them is the gold standard.鈥