New AI solution for smarter urban and climate planning

New AI solution for smarter urban and climate planning

Satellite images and other environmental data are stored and available as images. Through his solution Ontoraster, Dr Arka Ghosh has created an intelligent network in which every data point - every pixel in the images - is given a clear meaning and linked to other relevant elements.

"Think of it as a super-organised spider's web where every thread leads to the right information. Or in Minecraft terms: every block finally knows why it exists," says Arka Ghosh.

Faster decisions in crises and planning

By combining knowledge graphs with advanced AI, the system can understand how different data sources are connected. This allows users to ask complex questions and receive direct, comprehensible answers - without needing any programming or data analysis skills themselves.
The technology can be used in a range of socially important contexts.

"Urban planners can be supported in designing more sustainable and climate-smart cities. Companies can choose better locations. And authorities can act more quickly during, for example, heatwaves or floods," Arka Ghosh explains.

One example of a question the system can answer directly is: "Where are the hottest areas near densely populated districts in Umeå or Stockholm?"

"What previously required extensive manual analysis can now be done in a matter of seconds - without large amounts of data needing to be moved or processed," says Arka Ghosh.

AI that unlocks complex data

A central part of the solution is a refined AI technique that translates everyday questions into something the computer can understand. "This means that even people without a technical background can use the system and receive advanced analyses in return," says Arka Ghosh.

Funded by WASP

The Department of Computing Science at Umeå University has a strong connection to WASP - The Wallenberg AI, Autonomous Systems and Software Program -Sweden's largest individual research initiative aimed at positioning the country as a leading nation in the field. Several of the department's doctoral students, including Arka Ghosh, are funded through the programme, giving the research environment access to both national collaborations and cutting-edge expertise in AI.

AI-driven question-answering system

For several years, Arka Ghosh has immersed himself in international research on geodata, raster data, knowledge graphs and semantic technologies. The result is an AI-driven question-answering system that can interpret complex data in an entirely new way and provide meaningful answers to advanced questions. In practice, this means that authorities, engineers and other decision-makers can quickly obtain insights that previously required extensive manual analysis.

"I have tried to transform a chaotic pile of pixels into something that can actually be built with. A bit like in Minecraft: when the right blocks are put together, you can create something fantastic - but here it's about shaping the cities of the future and our environmental planning," he says.

Further information

Read the doctoral thesis here: "Semantic integration and query answering of multidimensional data with knowledge graphs"
Supervisor: Professor Diego Calvanese

Contact information