Geoscientists aim to improve human security through planet-scale POI modeling

Newswise — When geoinformatics engineering researchers at the Department of Energy’s Oak Ridge National Laboratory wanted to better understand changes in land areas and points of interest around the world, they turned to the locals — their data, at least.

Through an intelligent combination of geotagged social media, global location and natural language data, ORNL’s Junchuan Fan and Gautam Thakur developed MapSpace, a publicly available, scalable land-use modeling framework. By providing data characteristics broader and deeper than satellite imagery alone, MapSpace can generate population analytics invaluable for urban planning and disaster response. The researchers’ findings were published in the International Journal of Digital Earth.

“While the land cover — land, sea, water or desert — doesn’t change, the use of that land changes all the time,” said Thakur, leader of ORNL’s Location Intelligence group. “Understanding how land-use patterns change is essential for development of new services.”

Thakur said challenges with traditional satellite data are twofold. First, processing data accurately is an enormous task because of long download times and requirements of high-resolution imagery. Second, satellite images show only the tops of structures rather than their facades, which are more familiar vantage points. By using points of interest and geosocial data, Thakur and Fan can achieve multiple levels of semantic granularity, a measure of how accurately land use can be characterized.

For example, MapSpace could allow users to see that an area is commercial, zoom down to see it’s used for restaurants, then zoom in further to see the specific type of food served at a restaurant in the area.