City walkers are not computers
Smartphones have become omnipresent in our lives and one area in which we often can no longer survive without them is when navigating around cities. But does the optimised route offered by your maps app match up to the route that your brain would otherwise make you take to reach your destination? Based on research published in Nature Computational Science, the answer might be no...
Previous research has shown that people rely on landmarks and often miscalculate the lengths of streets when navigating in cities. Additionally, animals tend to navigate using vector-based navigation, whereby they tend to continuously point towards their destination, regardless of the obstacles in their way. Thus, taking the ‘pointiest path’ means constantly minimising the angle between your direction and the direction of your destination.
Using a large dataset of anonymised walking paths, Carlo Ratti and his colleagues at MIT showed that humans also take the pointiest path to their destination, rather than the shortest path. To their surprise, they found that this happened both within the grid-based structure of San Francisco, which is typical of many US cities, as well as in the more complex street layouts of Boston, which is similar to many European cities.
The reason for this is likely to be linked to minimising the strain on the brain. Evolution prioritises strategies that are good enough and use brain power efficiently, rather than spending extra energy to find the absolute optimal approach. Ratti speculates that thousands of years ago in the savanna using the pointiest path “would free up some additional computing power in our brains to avoid the lion on the path. Today we might do the same thing in the middle of the urban jungle to avoid the aggressive SUV.”
The study also demonstrated that the differences in time between following the pointiest path and the optimal path are not very large, and thus the pointiest path is often sufficient. However, the time lost for a particular route also depends on the urban grid.
Furthermore, they showed that humans tend to take different paths when travelling from A to B compared to when going from B to A. This is something that Ratti first came across when he was a postgraduate student at the University of Cambridge, “I realized that the way for me to walk from the college to the department was different from the way back... that was quite frustrating. I could not figure out if it was only me or if [it] was something I shared with everybody else in the city.” Thankfully, 20 years later, with the influx of data provided by smartphones, Ratti was able to get an answer to his ponderings: it turns out we all navigate cities in a similar way.
They filtered the large dataset of walking routes to isolate only the cases where people were walking between two defined destinations, without any navigation aid, and then compared these against various models employing different navigation approaches. They found that the pointiest path model had the closest match with the real life data.
Ratti is confident that the results could be validated in other cities and hopes that others can help to validate these findings based on data from around the world. He believes that if he had followed a more optimal path during his postgraduate studies in Cambridge he could have gained 5% of the walking time, which he could have used to work on his research, “I could tell my supervisor at the time that I would have finished my PhD much earlier.”