StreetLight Data is Bringing Big Data to the Transportation Industry
How would your forecasting and analysis improve if you could aggregate the historical location data of every car? What if cell phone data was also aggregated so you could analyze mode splits? What if we knew the purpose of each trip?
I’ve been blogging about AirSage for five years now, who promises to extract anonymous origin-destination matrices based on data from certain cell phone providers. My dream was that we could tap into their data stream to get much better origin-destination data for transportation planning studies as well as trip distribution patterns for traffic impact studies. One issue with AirSage’s data is that it is tied to cell phone towers, which limits the accuracy of the cell phone location. Another issue is that each project is bespoke meaning you need to work with AirSage who extracts the data for you.
StreetLight Data is a new player in bringing big data to our industry and it seems they are improving the AirSage concept. They have built a simple to use DIY system that allows us to:
- Quickly import or draw TAZs (traffic analysis zones)
- Choose the days and times for the data we want to extract (going back two years)
- Choose to extract origin-destination tables for the TAZs
- Choose a breakdown of trip purpose in the origin-destination tables
- Alternately, choose to extract travel times along a corridor by simply placing “gates” on a map
- See the price of the data we are requesting before we accept it
Here’s a snapshot from one of their case studies:
StreetLight Data has been working with Inrix to provide this data, which improves the location accuracy to about 5 meters. That’s interesting, but not necessarily a game changer over AirSage even though the car location data is about 10x more accurate.
The DIY system and 5 meter accuracy are really important features of StreetLight Data, but what is truly intriguing is that they have recently partnered with Cuebiq. Cuebiq provides the “location services” code that many cell phone apps use. This means StreetLight Data’s dataset will include anonymous location data from 35 million cell phone users.
Location data from 35 million plus devices gives them the penetration needed to provide solid origin-destination results in most parts of the country. It also allows them to build more interesting algorithms around trip purpose beyond home to office or home to retail.
I see a future using this big data where we’ll:
- Have accurate travel time data along any corridor (no more interns on out doing travel time runs)
- Be able to extrapolate pass-by and internal trip generation rates for every land use
- Have accurate data on mode share in our study areas
- Have accurate data on origin-destination patterns within our study areas
- Have accurate data on trip purpose in our study areas
The accuracy of our traffic forecasting will improve significantly with our use of Big Data. This is increasingly important as technology is changing transportation.
I’ll be blogging more about StreetLight as we use their service. But in the meantime, I encourage you to check out their site and watch their demo videos. Their service should significantly improve transportation planning.