To validate the results of their research, the scientists from MIT and Ford's Palo Alto Research and Innovation Center compared it to the model currently used by Boston's MPO. "The two models accorded very well," the researchers said in a paper published in latest issue of the Proceedings of the National Academy of Sciences.
"Mobile phones are the prevalent communication tools of the twenty-first century, with the worldwide coverage up to 96% of the population," the researchers said. "Mobile phone data have been useful so far to improve our knowledge on human mobility at unprecedented scale, informing us about the frequency and the number of visited locations over long term observations, daily mobility networks of individuals, and the distribution of trip distances."
Snapshot of the urban mobility simulation by the TimeGeo model at 6pm, bird's-eye view of Boston. For visualization purposes, only 1% of the 3.5 million individuals living in the Boston Metro are displayed in the video.
While the sparse nature of mobile phone usage leads to samplings that tend to have biases because they don't offer complete journeys in space and time for each individual, the researchers were able to infer certain mobility patterns.
For example, their algorithm assumes the location from which a user departs in the morning and to which he or she returns at night is home. It also infers that the location of the longest recurring stays during weekday daytime hours is the user's workplace.
The algorithm assumes most people's workdays are in accordance with national averages, so if a user makes phone calls from work only between the hours of 12 p.m. and 2 p.m., the system does not interpret that as evidence of a two-hour workday — unless that interpretation is corroborated by other data, such as regular calls from home at 11:30 a.m. and 2:30 p.m.
Signs directing pedestrian and street traffic in downtown Boston.
Any locations other than work and home are treated alike. From the available data, the system builds a probabilistic mobility model for each user, breaking every day of the week into 10-minute increments. For each increment, the model indicates the likeliness of a location change, possible destinations and the amount of time likely to be spent at each destination. The system then generalizes those probabilities across communities, on the basis of census data, and deduces cumulative traffic flows from the resulting probability map.
"We are able to identify home locations for 1.44 million users which is 75% of our initial user base," the researchers said. "Next, we filter users who have more than 50 total stays and at least 10 home stays in the observation period."
Sign up for CIO Asia eNewsletters.