How Automated Building Analytics Can Help African Cities Prepare for Climate Change

Jun 11 · 4 min read
Map showing individual building detections; the darker buildings are the most recent. The very light pink buildings did not change during the study period. // Map by Leanne Abraham, Planet

Climate change affects everyone, but it doesn’t affect everyone equally. It’s possible that satellite data and AI can help us better serve those most likely to feel its impact.

Two months ago, Cyclone Idai tore through Mozambique, causing floods that killed over a thousand people, leaving 400,000 homeless near the port city of Beira. In April, at least 60 humans died in KwaZulu-Natal Province, when torrential downpours ripped sinkholes in roads, burst rivers and crumbled buildings. In 2017, Liberia and Sierra Leone endured mudslides and flash floods that resulted in the deaths of hundreds.

This is just a small part of a much larger story of the effects of climate change on Africa. According to the UN, a third of Africans live in drought-prone regions, where a drying climate could put the lives and livelihoods of an additional 75–250 million people at risk by 2030. Across southern Africa, flood-prone areas are becoming wetter as rainfall patterns shift, causing floods and cyclones to become more frequent and severe.

Moved by the urgency of these trends, engineers at Planet are researching new ways that satellite imagery and analytics could help us understand these shifting climate risks.

Using Planet’s path-breaking automated building detection analytics, our engineers conducted a study (full conference paper here) that tracked urban growth in five African cities built on flood plains, which made these cities uniquely vulnerable to climate-enhanced flood events.

Map of study site locations // Map by Leanne Abraham, Planet

According to the study, in the city of Bangui in the Central African Republic, 57 percent of the detected buildings were in a high-risk flood zone, and there was a 2.5 percent growth in that zone during the monitoring period from July 2017 to March 2019. In Bamako, Mali, 18.7 percent of detected buildings were in a risk zone. In Casablanca, Morocco, 13.7 percent were in a risk zone. The numbers are smaller but still significant in two other cities that the team studied: 7.0 percent in Ouagadougu, Burkina Faso, and 2.1 percent in Addis Ababa, Ethiopia. Notably, in all five cities — building growth in high-risk flood zones was increasing.

“We didn’t expect to see numbers that high,” says Ramesh Nair, study co-author and computer vision engineer on the analytics team at Planet. “We’re excited that Planet’s tools can help us see and better understand the flood-risk issue.”

Planet’s imagery and automated building detection algorithms are a critical resource for flood-risk specialists, helping them build models with unprecedented accuracy and geographic coverage. These models can help urban planners, the private sector, and municipal and other government leaders to develop new disaster-risk-reduction strategies, ranging from new climate-smart infrastructure investments to new insurance mechanisms.

“To reduce the effects of flooding in fast-growing cities, urban planners need up-to-date environmental sensing data and analysis over large geographic scales,” says Gopal Erinjippurath, study co-author and senior director of analytics engineering at Planet. “In the developing world, such data just hasn’t existed until now.”

“Climate change is a ‘super-wicked’ problem — a nested tangle of complex social and ecological challenges. To build resilience, we need an ability to see those challenges from many angles,” says Andrew Zolli, vice president of global impact initiatives at Planet. “By imaging the Earth’s landmass every day, and enabling analytics that can help us understand many forms of connected change, Planet helps do just that.”

Overview map showing areas where there was a large amount of building-growth detected // Map by Leanne Abraham, Planet

Planet has previously worked with the World Bank to pioneer new tools to detect urban change, beginning with Dar es Salaam, Tanzania. The project falls under the Tanzania Urban Resilience Program.

“It’s exciting to work at the forefront of both space and machine learning, and the power of the infrastructure we’ve built to automate this analysis is awe-inspiring,” says Christian Clough, study co-author and machine learning engineer at Planet. “I’m grateful that our data can be used to help people.”

To connect further with the study’s authors, reach out to Christian Clough:

Planet Stories

Using space imagery to tell stories about our changing planet.


Written by



Planet Stories

Using space imagery to tell stories about our changing planet.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade