Complete Trip: A Linked Multimodal Human Mobility Dataset
This paper introduces "Complete Trip," a novel dataset that reconstructs linked multimodal human mobility journeys from smartphone location data across six Utah counties in 2020, offering network-based route representations and population-level inference weights to support reproducible research in transportation, public health, and urban science.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine the world as a giant, invisible dance floor where billions of people are constantly moving. Scientists have long wanted to understand the choreography of this dance: Where do people go? How do they get there? Do they walk, drive, or take the bus? And most importantly, how do these individual steps connect to form a full story? For years, researchers only had pieces of the puzzle. Some data showed where people were at a single moment, like a snapshot. Others showed where they started and stopped, like a start and finish line, but missed the messy middle. Some data was just a blurry crowd count, while other data was so private it couldn't be used at all. Without the full picture, it's hard to plan better cities, stop the spread of diseases, or help people get to work on time. We needed a way to see the entire journey, not just the dots on a map.
This is where a team of researchers from Villanova University steps in with a new tool called Complete Trip. Think of this dataset as a magical movie reel that stitches together thousands of tiny, scattered video clips of people's lives into one smooth, continuous film. Instead of just seeing a person appear at a coffee shop and then disappear at a park, this dataset shows the whole story: the walk to the bus stop, the ride on the train, the transfer to a different bus, and the final walk to the park. It does this by taking "passive" data—signals sent by smartphones when people have location services turned on—and cleaning them up to reveal the full journey.
The paper presents this new dataset, which covers six counties in Utah for the entire year of 2020. It's like having a time machine that lets us watch how 14.8 million anonymous devices moved around during a very specific year. The researchers didn't just collect the raw signals; they built a four-step factory to turn those signals into useful stories. First, they figured out when a person started a trip and when they stopped. Second, they guessed how they traveled (car, bus, train, or walking) using a smart computer program. Third, they drew the exact path they took on a digital map, rather than just guessing the straight line between two points. Finally, they linked these individual steps together. If a person took a bus to a train station and then switched to a train, the dataset knows these are part of the same "journey," not two separate, unrelated events.
What makes this special is that it connects the dots. Previous datasets were like having a list of addresses people visited but no idea how they got there, or a list of bus rides without knowing who was on them. This new dataset links everything. It shows that a trip isn't just a point A to point B; it's a chain of events. The researchers also added a special "magic number" (called a weight) to each trip. Since smartphone data doesn't capture everyone equally (for example, it might miss older people or those without smartphones), this number helps scientists adjust the data so it represents the whole population, not just the phone owners.
The team tested their work to make sure it was real. They compared their data against official traffic counts and transit records. They found that their data matched the real-world trends very closely, with a correlation of 0.96 for bus and train ridership after applying their adjustments. This means the dataset is a reliable mirror of what actually happened on the roads and rails in 2020. They also made sure to protect people's privacy. Instead of giving exact addresses, they turned locations into fuzzy grid squares (called Geohash-6), and instead of exact times, they used 30-minute blocks. You can't see exactly who went where, but you can see exactly how the crowd moved.
In short, this paper doesn't just give us a list of numbers; it gives us a new way to see human movement. It suggests that by linking trips together and mapping them onto real roads and rails, we can finally understand the full complexity of how people live and move. Whether it's figuring out why a bus route is empty, planning where to build a new bike lane, or understanding how a city reacts to a storm, this dataset offers a clearer, more complete view of the dance floor. It's a tool that turns scattered, confusing signals into a coherent story of how we get from here to there.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.