Smart Transportation Startup Uses AI and Geospatial Data
Last year, we got a glimpse of what the world would look like without us – no crowds, no traffic, no nonsense. That led to the biggest drop in greenhouse gas emissions since World War II. Of course, all good things must end. Traffic in many areas is returning to pre-pandemic levels or even higher, though some patterns may never return to normal. Sitting bumper to bumper on the highway isn’t just bad for the environment and your mental health. It also affects the economy, with congestion causing $87 billion in lost productivity in just the United States in 2018 alone.
Countries like China have turned to artificial intelligence to manage traffic, helping speed things up by at least 11% in cities such as Hangzhou, formerly one of the most congested in the country. Plenty of others are adopting smart transportation technologies like that offered by startup StreetLight Data.
Geospatial Data for Smart Transportation
We’ve written quite a bit about the value and applications for what’s called geospatial intelligence, often in the context of using satellites and AI to glean profitable insights about economic activity. The simplest use case is the old counting-the-cars-in-the-mall-parking-lot scenario to gauge the retail environment. Since there are no more malls, that application probably isn’t as relevant as it used to be. However, the concept of geospatial intelligence can apply to more than satellite or aerial imagery, as it refers to analyses of any geographic-based data, such as tracking cell phone locations. We like to refer to this as geospatial data, and it’s this kind of information that StreetLight Data leverages to offer what it calls the world’s only interactive transportation data platform.
Founded in 2012, the San Francisco startup has raised $44.4 million, including a $15 million Series D last summer. We first came across StreetLight Data nearly five years ago in our article on big data companies watching your every move. In the case of StreetLight Data, it ingests, indexes, and processes about 40 billion anonymized location records from smartphones and navigation devices in connected cars and trucks every month. The company’s machine learning engine, Route Science, combines that location data with other information, such as embedded road sensors, to provide a variety of insights on traffic patterns that are available on its StreetLight InSight web platform.
Counting Cars for Smart Transportation Management
One of the company’s flagship products is its Average Annual Daily Traffic (AADT) tracking tool that covers 4.5 million miles of roadway in the United States and Canada, based on more than one trillion location records. StreetLight validates its AADT data against nearly 10,700 permanent counters across the United States and another 1,300 permanent counters across Canada.
In the old days (and still), traffic planners would physically deploy road tubes to collect information about vehicle volume by sensing air pulses that are created by each tire strike on the roadway. A second tube can collect additional details such as speed. Streetlight Insight can do all of that and much more digitally, including average trip speed, duration, and distance, as well as detecting major concentrations of commuters and shoppers. And it’s cheaper:
Let’s take a look at a few use cases to understand the wide range of applications for the technology.
Smart Transportation Use Cases for StreetLight InSight
The thing to keep in mind is that StreetLight provides more than just metrics – how many cars speed through school safety zones, for instance – by drilling deep into the data to provide additional context.
In Florida, for example, state officials wanted to know how much of the heavy truck traffic on the I-75 corridor was short-haul versus long-haul vehicles. Doing what the company calls an Origin-Destination (O-D) analysis, StreetLight InSight determined that a higher percentage of southbound truck trips were long-haul, which provides hard data for moving some of that traffic to other routes to keep the main highway artery clear for commuters and sightings of Florida Man.
In a different case study, a small city wanted to use StreetLight to determine whether its efforts to prevent cut-through traffic from leaving the highway and clogging local streets had any effect. First, an O-D analysis validated the city’s suspicions that commuters were indeed taking a shortcut through town. Next, city planners measured the “before and after” effect of new measures such as turn restrictions and a ramp meter. It turns out that the government isn’t always ineffective, as the analysis found the planners actually knew what they were doing:
- 25% more trips stayed on the highway during peak hours
- Cut-through traffic on four highway on-ramps decreased by more than 25%
- Volume on local roads decreased by 33% overall
The city ran more than 35 analyses for one fixed fee, with an average analysis run-time of eight minutes. “This saved thousands of dollars (and untold hours) compared to outfitting the same scope of the city with sensors,” according to StreetLight Data.
Driving Smart Transportation Analytics
StreetLight Data claims that it supports more than 6,000 smart transportation management projects each month, ranging from infrastructure planning to accommodate more electric vehicles or hubs for urban air mobility to studies with a more socioeconomic angle, such as ensuring equitable public transportation options and even reducing infant mortality rates. The company is obviously trying to be a thought leader in the space, releasing its own data reports, especially on the ongoing impacts of COVID-19. For example, in a recent report, StreetLight Data noted that the precipitous drop in total vehicle miles traveled (VMT) in 2020 did not strongly correlate to the decline in gross domestic product (GDP) last year:
It also captured how driving patterns continued to evolve during the pandemic, especially as VMT rebounded through the year. Specifically, commuters could zip to work in the mornings, but faced a longer stretch of after-work traffic in the afternoons:
The payoff is turning all of this data into actionable insights. For example, StreetLight Data says companies could adopt a once-a-week work-from-home policy that would cut VMT by 20% without affecting economic output. In fact, as we discussed earlier, congestion actually costs billions in lost productivity. That’s why you’ll never find us commuting to an office. Plus, we don’t really have one.
StreetLight Data is one of the original big data companies for smart transportation intelligence. The company is tackling a broad range of use cases, from traffic planning and engineering to smart city solutions including first- and last-mile delivery, and everything in-between. More importantly, the Bay Area startup is showing ROI across its business portfolio.
StreetLight Data isn’t the only company leveraging smartphone location data. Previously, we wrote about SafeGraph, which uses AI to process similar geospatial data, but focused on retail and marketing using foot traffic. And, as we mentioned earlier, all kinds of satellite startups are using aerial imagery for geospatial analysis. These and other types of alternative data will become increasingly important for businesses in need of a competitive edge. Big tech companies like Google would certainly be interested in some of these alternative datasets, as its digital advertising business comes under more and more government scrutiny.
Tech investing is extremely risky. Minimize your risk with our stock research, investment tools, and portfolios, and find out which tech stocks you should avoid. Become a Nanalyze Premium member and find out today!