You are probably not a traffic engineer, but now you very well may be assisting the Flagstaff Metropolitan Planning Organization in mapping and modeling traffic patterns in Flagstaff and across the county, all by just carrying your smartphone on your daily commute.
Big data has hit Flagstaff as FMPO is now using cell phone data to help better understand the city's traffic.
The data, purchased from the San Francisco-based company StreetLight Data, is gathered from smartphone applications that use location services.
After being downloaded onto a smartphone, many apps will ask the user for permission to access the phone’s location. If approved, the app will then be able to ping a user's movements, creating a mountain of location data each day.
After being used by the app, the location data doesn’t just disappear. Instead companies sell it, often to advertisers who can use it to sell products and services nearby.
However, that data is also being acquired by FMPO said David Wessel, the organization’s manager. It has already begun to change the way they see transportation in Flagstaff.
It does not, however, change the way they see any one individual in the city. When the data arrives, it has already been made anonymous, Wessel said, with much of it already having been turned into a map format.
In fact there are some areas, mainly in rural parts of the county, in which they might not be able to use the data because so few people live there and identification could become likely.
As the CEO of the Northern Arizona Intergovernmental Public Transportation Authority Erika Mazza said, “We could see that there might be 75 people who traveled from Tempe to Snowbowl one day, [but that’s all].”
Nonetheless, the data could change the way the city tries to manage traffic, give staff a better understanding of how developments or new roads may create or reduce congestion and change the routes buses take.
For example, Wessel said through the data, they quickly found original estimates for traffic patterns were now off. The original estimates for traffic had been found during a study in the late 80s.
At that time, such data was collected by posting people outside Flagstaff. They would then try to wave down passing vehicles before asking the occupants to fill out a survey on where they were traveling to and from.
As one might imagine, Wessel said, this was not easy and they would be lucky to get a sample size of even 1 percent.
A similar study was repeated in 1998 when FMPO staff sat at truck stops an hour or so outside Flagstaff and asked those who stopped where they were traveling to and from.
This also may have skewed the data, Wessel said, as the majority of people who were driving after just leaving Flagstaff, or only an hour away from their destination in the city, were probably less likely to stop than those in the middle of a trip.
Now capturing about 18 percent of the trips within and across the county, the cell phone data gives the FMPO a far more accurate picture of traffic.
You have free articles remaining.
TRAFFIC ON MILTON
Some aspects of what the data shows are more obvious. When it comes to those traveling to and from Northern Arizona University every day, the data shows that a much larger percentage of traffic coming from sections of the city that contain more apartment buildings.
At the same time, they also found that very few people traveling to and from NAU use Milton Road with far more simply crossing the street.
On Milton, it also shows that 60 percent of the traffic on the road doesn’t actually travel along the whole stretch. Instead the majority of people who travel on Milton have a destination that is somewhere along the road itself.
Being able to see patterns like this may influence the traffic solutions that are found for Milton, especially as the Arizona Department of Transportation is in the midst of creating a plan for the future of the road.
For the bus system, this data can also be useful in a number of ways, said NAIPTA development director Kate Morley.
“This is our first time using this kind of data and we're really excited,” Morley said.
For one, Morley said, the data can provide travel time. Passengers and residents often are concerned about congestion on roads, but beyond simply counting the number of vehicles, congestion can actually be hard to quantify and measure.
Travel time can be a good way to do so, with it increasing when roads are congested and decreasing when they are not, Morley said. It can also help them by providing a clue as to how well they are providing service to the city.
By laying their bus routes over the data maps, Morley said they can tell if the routes they are providing are bringing people to where they are going in an efficient manner or if they may need to change. Moving forward, the data can show them where new routes may be best placed to reach populations they don’t already serve.
Mazza pointed to the Ponderosa Trails neighborhood. Looking at the data, they noticed that Ponderosa Trails has a high travel rate to the Flagstaff Medical Center, likely because many people who work at FMC may live in that neighborhood.
So looking to the future, Mazza said they could implement a bus route that more directly ties those two areas together.
Wessel said they are also hoping to compare the traffic data to the demographics of the city. By doing this, they may be able to compare the data of travel patterns to data on income as a way to see how and where lower income residents, who may have lesser access to cars, travel most.
This could influence how the city implements future transportation systems, perhaps providing better public transportation to these areas, Wessel said.
However, this has not come cheap and Wessel said they only have access to the data until mid-March and they don’t know whether they will buy the data next year.
Instead, they are using it to help build a better model of the city's, and parts of the county’s, traffic. By calibrating the model with this data, and using the more traditional sources of data they have such as simple traffic counts, they will be able to reap the benefits long after the data is no longer available.