Now that face masks and social distancing have become the by-default behavior for anyone stepping into a public space, many former commuters will find the very thought of jam-packed train carriages at 5.30 pm on a weekday a hair-rising memory. But as lockdown regulations lift and commuters start using trains and buses again, local councils are having to find smart ways to monitor human flows on a daily basis.
And in an attempt to make rush-hour public transportation a thing of the past, the UK's train and coach app Trainline has added a new crowdsourced feature to its service, to warn passengers if their journey is looking busy.
Commuters across the nation who use the Trainline app now have the option to participate in Crowd Alerts, to warn fellow passengers if it is not possible to social distance on a train or a specific carriage.
SEE: Guide to Becoming a Digital Transformation Champion (TechRepublic Premium)
A "live tracker" button has appeared in the app for any upcoming journey, which, when tapped, asks the user to confirm whether social distancing is possible on their train by clicking a "thumbs up" or "thumbs down" option. If it is a thumbs down, the passenger will be asked whether they are at the front, middle or back of the train, and the feedback will be sent to other users of the app – an orange bubble for a busy train, and a blue one for a train that is only partially crowded.
The system is designed to avoid sneaky passengers falsely reporting a busy carriage just to make sure they get extra leg space. The bubbles indicate how many commuters have shared the information, and each user can only provide feedback for a number of minutes. Passengers are only asked about a particular section of one train; plus, any fresh feedback – hopefully from an honest user – will override previous reports.
"We have automated tools to make sure that, even if you had an army of minions reporting a particular carriage is busy, we'd be able to clean that up," Dave Slocombe, senior product director at Trainline, told ZDNet.
"The very worst that you could possibly do is feedback that one section of the train has poor social distancing, and falsely direct passengers to other carriages. The impact of that isn't hugely disruptive," he added.
The real-time data generated by Crowd Alert is also providing useful insights about the way people move in various areas around the country, which can be shared with the wider industry to help operators manage train lines and timetables as the country exits lockdown.
Since Crowd Alert launched a few weeks ago, commuters have reported that social distancing was feasible in about 90% of the trains running across the nation, suggesting that the government's advice not to travel at peak times is having an impact on people's behavior.
While crowd numbers used to spike between seven and nine in the morning, and then again at the end of the working day, Slocombe said the new feature shows that passengers are now travelling at alternative times, and rarely in the evenings.
"The busiest time of the day is probably around midday, but generally there is an even, flat curve of people travelling," said Slocombe. "That means that social distancing is effectively being observed."
Of course, crowdsourcing information to anticipate the "busyness" of public transportation is nothing new. In fact, Slocombe explained that the Trainline team built Crowd Alerts on the backbone of a tool called BusyBot, which the company was using prior to the pandemic to inform passengers of the likelihood that they would find a seat on their train.
When BusyBot was live, almost 30,000 passengers told the tool whether seats were available in their carriage every day. The data powered a predictive model, which then anticipated whether future trains would be busy.
SEE: Can a smarter app get you out of your car and onto a train instead?
The system is similar to Google Maps' transit crowdedness predictions introduced last year, which leverage details provided by passengers on past bus, train or tube rides to estimate how crowded a given commute will be.
Predictive algorithms, however, were based on trends in public transportation that a few months of lockdown have rendered completely obsolete. "The predictive model we used for BusyBot stopped working when timetables changed and much less people traveled," said Slocombe. "So instead, we connected the crowdsourced feedback to a real-time system."
Similarly, Google Maps is now letting users voluntarily opt in to Google Location History when they are travelling, in order to gather anonymized data that reflects how crowded transportation networks are.
The search giant is welcoming contributions from passengers, to inform a live data board in the Maps app that shows how busy certain stations or lines are compared to usual levels of activity – although the service, unlike Crowd Alert, does not specifically show whether social distancing is still possible in the areas covered by Maps.
Predictive technologies, therefore, are now having to re-create datasets that are in line with new realities – and it means going back to their very first, crowd-sourced version.
Slocombe explained that commuters seem keen to provide feedback: up to 70% of trains in the UK are covered with user-based information thanks to the new Crowd Alert feature. Yet the Trainline product director stressed that developers tend to prefer eventually moving away from crowdsourcing.
"BusyBot will be coming back once CrowdAlerts isn't needed anymore," he said. "In a time of high change, real-time feedback is great. But we like the predictive model when things are more predictable, because it means we can ask customers to do less work."
Crowd Alert is available across the UK in beta, while the Trainline team refine the feature's last details. But it seems that being squashed against fellow passengers on a sweaty summer commute needn't be a part of your daily routine for the foreseeable future.