State uses analytics to build roads faster, cheaper
- By Rutrell Yasin
- Nov 28, 2011
North Carolina’s Transportation Department is using analytics software to build roads faster and for less money while minimizing environmental disruption.
NCDOT is analyzing geographic data to help narrow the choices of possible road corridors and, at the same time, is able to reduce costly land surveys. The process can save $500,000 per project and shave 20 percent off the time needed to select and plan a road, according to North Carolina officials.
NCDOT and the North Carolina Division of Water Quality are working collaboratively on the project, using SAS Analytics as the engine to analyze volumes of geographic data.
How data wizardry can revive America’s cities
Planners need to protect water sources as they assess transportation projects. As a result, they have to propose ways to avoid or minimize environmental impacts. Road builders often verify geographical surveys by sending surveyors and water quality experts into the field to document streams and wetlands. However, a large project, such as a bypass, requires substantial time and manpower to survey thousands of acres in order to identify environmental issues.
“You might have hundreds of possible combinations for one road,’’ said Morgan Weatherford, environmental program consultant with NCDOT’s Natural Environmental Section. “It’s a major challenge to comply with federal and state environmental regulations in a manner that is beneficial for the environment and taxpayers as well.”
A new data source has emerged in recent years with the potential to eliminate some costly field work. Light detection and ranging (LIDAR), which uses laser pulses to record the distance between two points, is useful for charting land elevation, which is key to locating wetlands and streams. LIDAR data is used extensively to update flood maps and is considered more detailed than geological survey information.
However, LIDAR produces large volumes of data. One transportation project might involve upward of 30 million records with 30 different attributes per record. Prior to using SAS analytic software, nobody at NCDOT had used LIDAR data to predict stream and wetland locations for construction planning purposes.
The NC Division of Water Quality built models to predict headwater streams and tested the accuracy with field surveys.“The models were 85 to 95 percent accurate, depending on the terrain,’’ said Periann Russell, environmental senior specialist with the NC Division of Water Quality.
NCDOT also used LIDAR data for a much larger project that includes predicting stream and wetland locations for an entire county. The data will help transportation planners choose a corridor for a 20-mile bypass. It will also be used on several bridge modernization projects and be available to private developers for their proposed developments, NC officials said.
Rutrell Yasin is is a freelance technology writer for GCN.