
Junchuan Fan and Gautam Thakur helped develop MapSpace, a land-use modeling framework developed from natural language data, geotagging and global location data. Credit: Lena Shoemaker/ORNL, US Dept. of Energy
When geoinformatics engineering researchers at the Department of Energy’s Oak Ridge National Laboratory want to better understand changes in land areas and points of interest (POIs) around the world, they turn to locales—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 have created MapSpace, a publicly available, scalable land-use modeling framework. By providing data characteristics that are broader and deeper than satellite imagery alone, MapSpace can generate population analytics valuable for urban planning and disaster response. The researchers’ findings were published in International Journal of Digital Earth.
“While land cover — land, sea, water or desert — doesn’t change, the use of that land changes all the time,” said Thakur, head of ORNL’s Location Intelligence group. “Understanding how land use patterns are changing is essential for developing new services.”
Thakur says the challenges with traditional satellite data are twofold. First, accurate data processing is a huge task due to long download times and high-resolution image requirements. Second, satellite images only show the tops of structures instead of their facades, which are more familiar points of view. 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 described.
For example, MapSpace may allow users to see that an area is commercial, zoom down to see that it is used for restaurants, then zoom in to see the specific type of food served at a restaurant in the area.
The ability to capture land use characteristics includes not only how buildings and areas are used but also how their function can change over weeks or even a day.
“Land use layers are traditionally static, but there are different places in the world where land use changes during the day,” such as a town square used as a farmer’s market on weekend mornings, Thakur said. “How to capture these dynamics in spaces and how to change them is also one of the things that Junchuan is leading.”
By understanding change over time, researchers can propose new purposes for land use that may not have been possible before. For example, vacant urban areas that once housed thriving shopping malls may be better used for the development of apartments or industrial centers.
Perhaps almost as impressive as the capabilities they have developed is how quickly Thakur and Fan can do it in one geographic area. Fan was able to develop land use data for the entire African continent in just two weeks. Before MapSpace, it would take months, if not years, he said.
The speed of the program is greatly expanding its reach in areas such as national and human security. Fan gives examples of emergency support after a hurricane or responding to a public health crisis as the need for rapid land use mapping. MapSpace data can provide insight into safe places to send emergency vehicles or the most accessible locations to set up aid stations.
“For our research to be able to produce global-scale data and map that in real time—or in a short time—that will help us better respond to some time-critical missions,” Fan said. “That’s the unique dimension of the lab’s work. Here, it’s important to have both a scientifically sound approach as well as a scalable data product that can be easily deployed.
Because of its diverse resources, ORNL is well suited to lead a project like MapSpace. Thakur credits the researchers’ success to the interdisciplinary expertise available at the lab and the world’s unparalleled computing resources that enable rapid, geographically sound land-use data on a global scale. Many of the lab’s experts in data science, geographic information systems and geography contributed to the project. Thakur also noted that the lab’s study of national security provides a unique perspective for this type of research and development.
“Combining expertise in different domains and creating an interdependent research capability is something unique to ORNL that we rarely find in industry or academia,” he said.
The fan continues to map different continents; his next project is South America. He started with areas outside the United States because they had a scarcity of data and would benefit from a valuable source that could fill in the gaps in land use. With the capabilities developed by Fan, Thakur said it is possible for the team to map the entire planet every year, with increased accuracy and spatial and temporal granularity.
“I don’t know anywhere in the world where that is possible at this time … even on a national scale,” Thakur said. “I think the capabilities we’ve built here in the lab are not only unique but also first [of their kind] in the world.”
More information:
Junchuan Fan et al, Toward POI-based large-scale land use modeling: spatial scale, semantic granularity, and geographic context, International Journal of Digital Earth (2023). DOI: 10.1080/17538947.2023.2174607
Provided by Oak Ridge National Laboratory
Citation: Geoscientists aim to improve human security through planet-scale POI modeling (2023, July 25) retrieved on July 25, 2023 from https://phys.org/news/2023-07-geoscientists-aim-human-planet-scale-poi.html
This document is subject to copyright. Except for any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. Content is provided for informational purposes only.