The product, Atom, can “learn” how to transform data on its own, and draw on Google’s proprietary traffic data taken from user’s mobile phones.
A government employee takes a photo of a wall with a tablet. Software on the tablet finds where the cracks are in the wall, and turns that into data that can be tracked over time.
Machine learning, meet the transportation agency.
That’s one of the larger selling points that SADA Systems, a software firm with close ties to Google, is pointing toward as it debuts an asset management platform called Atom this week at the annual meeting of the Transportation Research Board in Washington, D.C. The platform uses Google’s machine learning capabilities — as well as Google Maps and Waze — to add layers of service on top of the core concept of managing assets. And the target customer is state and local government transportation offices.
The machine learning component means more than just recognizing features such as cracks by taking photos. Atom, delivered in a software-as-a-service model, can “learn” how to transform data on its own so that neither the user entering data nor an administrating employee needs to change anything themselves. It can draw on Google’s proprietary traffic data taken from user’s mobile phones. And it can push out notifications to Waze users, letting them know about work crews on their routes.
“There are aspects of our solution that are similar to other products in the market, but there’s a lot that we do that’s pretty unique and proprietary to our system,” said Patrick Skoglund, director of geospatial technologies for SADA.
At the product's core is a system that can track assets, compile data about them, offer users in the field and the office a shared platform to collaborate through, allow managers to track and assign tasks to staff, and help deliver information to people who might otherwise not communicate with each other. For example, Skoglund suggested, if a utility needs to do work on a certain road, they could use Atom to see whether any other crews are doing work in the area and when. On the back end, transportation officials could then use Atom to figure out whether the crews could share costs or work together in any way.
It can also handle notifications for users working in the field.
“If an administrator wanted to go into our system and say, ‘X, Y and Z clearance heights are under 12 feet,’ we would notify the appropriate trucks of that clearance height,” he said.
The product itself has grown directly out of government use. According to Skoglund, Atom is the market version of a framework that the city of Chicago has been using for three years. All transportation work in the city flows through that framework, and Chicago has more than 700 users in the system to date. That framework has also plugged into the city’s open data efforts — particularly, WindyGrid.
That’s another big piece of the product: public information. Atom's public portal component can provide the data to help the public answer questions and learn about transportation, Skoglund said, and it gives users the ability to create custom data maps.
“If an administrator of Atom wanted to create an iframe for a website or even generate a URL," he said, "they can specifically call out what information they want to send to the public from a mapping perspective."