Google Open Buildings helping strengthen community resilience

Google says its artificial intelligence (AI) models are being trained to extract optimal information on building footprints from satellite imagery.

The search giant shared the impact of AI through its “Open Buildings” data set during a virtual briefing on Tuesday.

Fresh, on-the-ground evidence from health planners and climate analysts in Rwanda, Nigeria, and South Africa demonstrates how the Open Buildings dataset, which maps more than 1.8 billion structures, is helping frontline projects close service gaps and strengthen community resilience.

‘Open Buildings’

Google West Africa’s Communications and Public Affairs Manager, Taiwo Kola-Ogunlade, said the improvements made over time extend beyond just the building’s footprint.

“We are evolving to include building height and temporal data so you can see not just the place at one point in time, but across various years now, and this dataset is open today. We have a couple of stories with real-life organisations that have used this information to improve their operation

“Over 1.8 billion building detections across the global south have been in fact to date, and we have a dataset that stretches close to six years, from 2018 until last year, 2024, and all of this information is publicly available to date,” Kola-Ogunlade said.

ALSO READ: Google I/O 2025: More AI in Gmail, Gemini, Search and Android XR [VIDEO]

Real world

Kola-Ogunlade added that the “Open Buildings” data set has mostly been reactive and open.

“But we’re trying to be more intentional about telling the stories so that other organisations, countries, governments, entities, can figure out how to improve their planning, their execution and their service delivery”.

AI tools

Google said that across the African continent, AI tools are transitioning from pilots to public-service workhorses, powering everything from crop price forecasts to live language translation.

The search giant said Open Buildings—developed by researchers at Google Research Africa—is part of that shift, proving that open, locally guided AI can turn sparse data into timely decisions without locking African problem-solvers behind paywalls or proprietary gates.

“The World Resources Institute (WRI) has already applied this flood-risk modelling in Dire Dawa, Kigali, Musanze, Gqeberha and Johannesburg, combining two-dimensional hydraulic simulations with building footprints to create block-by-block exposure maps.

“The need is pressing: Nairobi is still recovering from the May 2024 floods that displaced more than 278 000 residents, while Johannesburg’s rapid growth along the flood-prone Jukskei River continues to heighten vulnerability,” Google said.

Africa

Google said the projects tackle challenges felt across the continent.

Africa still has about 8.7 million “zero-dose” children who receive no routine vaccines each year, while the urban population is projected to almost double to nearly one billion people by 2035, forcing new homes into high-risk floodplains and informal settlements.

“By giving ministries, city engineers and NGOs a street-level view of where communities are growing and how they change over time, the Open Buildings dataset lets limited budgets flow to the people and places that need them most,” Google said.

Expertise

Google Research Africa Research Scientist and Head, Dr Aisha Walcott-Bryant, said the projects demonstrate what happens when local expertise meets open, scalable technology.

“Mothers reach clinics sooner, children receive long-overdue vaccines, and city planners get ahead of the next flood. We’re thrilled to stand beside partners who turn data into decisive action, proving that AI’s highest calling is to solve real-world challenges.”

“From lifesaving clinics to climate-smart cities, these projects show how open AI is fast becoming Africa’s quiet engine for equitable progress,” Walcott-Bryant said.

The Open Buildings dataset—along with sample code, tutorials and its new 2.5D layer—is freely available at Open Buildings.

ALSO READ: Google launches cloud region in Johannesburg [VIDEO]