It’s long been feasible to simulate how individuals move through metro areas, although the source data for this analysis was mostly synthetic and not based on the activities of a real population. With the advent of smartphone location data, it’s also been possible to study the aggregate flow of millions of people from one area to another via their device location. Now, by integrating the two, AirSage is able to provide high-fidelity mobility insights on many hundreds of thousands, even millions of individuals, as they travel by roads and public transit.
Smartphone Location Big Data
Our smartphone location data is available for 150 Million+ devices each month, and provides billions of locations per day. Our data panel, predominantly from smartphone GPS, gives us extraordinary insight into the activity and overall movement of the entire country. By identifying the home neighborhood location of each device, we estimate the demographic distribution of our sample. We treat data privacy very seriously and have rigorous protocols for removing personal information from our sample; moreover, all our products are built utilizing aggregated data and we never provide data for individual devices.
Modeling the Movement of Millions of People
Using AirSage’s GPS data and a scalable simulation engine (akin to Google Maps on steroids) with nationwide road and transit networks, we can show the likely path of travel for an individual, moving from point A to point B. Because the simulation models millions of people simultaneously, we can also calculate the volumes, speeds, and travel times along each road segment, for any hour of the day. By including rail and public transit network data, we can also estimate the percentage of use for different travel modes (e.g. car, rail, bus) for a given population. This approach allows for studying patterns of population movement and what-if planning scenarios for an entire metropolitan area.
Mobility Analytics from Real-World Location Data
But how exactly do we do this? For a realistic model of population mobility, we use our smartphone location data to first create a device-level activity profile that is the input to the large-scale travel simulation. This ensures the population movement is based entirely on real-world observations. The simulation takes this initial travel demand and loads it onto the transportation network, to route the devices while constraining them to the road capacities and travel speed limits. This process is repeated over many iterations to estimate roadway link-specific volumes, speeds, and travel times. In addition to these aggregated travel statistics, we also provide:
- Demographic distributions (e.g. age, income, gender, education) on every roadway link
- Mobility insights based on lifestyle segmentation
- Impact of hypothetical population or travel network changes on user behavior
This allows us to answer questions such as:
- How are customers traveling to a point of interest, and how much time do they spend in travel?
- What are the demographics of people driving past a specific billboard during 4pm to 6pm, and how fast are they moving past it?
- What income groups are most affected by changes in the road network or in public transit?
Our nationwide mobility analytics will provide additional insight to all of our location intelligence products, and has wide applications in travel and tourism, transportation, advertising, and finance. We're excited to be introducing these new features in the coming months and look forward to engaging with customers to answer questions that were previously impossible to answer.