Transforming refugee camps with solar power and open-source mapping

Key Concepts

• Humanitarian OpenStreetMap (HOT): A collaborative project to create and provide open street maps for humanitarian action and response.
• AI for Good Lab: An organization developing open-source AI tools for social impact.
• Ground Truth Data: Manually verified data used to train and validate AI models.
• Drone Imagery: Aerial photographs and videos captured by drones, used for mapping and analysis.
• Open Source: Software or data freely available for use, modification, and distribution.
• Grid Extension: Expanding the reach of existing electricity networks.

Mapping Kakuma Refugee Camp: A Collaborative Approach with AI

The video details a collaborative effort to map the Kakuma refugee camp in Kenya, home to over 300,000 refugees from more than 20 countries. The primary goal was to improve logistical decision-making and identify opportunities for expanding access to electricity, ultimately supporting economic empowerment within the camp. Juan emphasizes the critical need for accurate maps, stating, “Like every city, having maps is critical for logistics decisions, understanding where to locate specific services. A lot of times we don’t have that information in refugee camps.”

The Challenge of Manual Mapping & The Role of Community Involvement

Traditionally, mapping areas like Kakuma manually is “incredibly prohibitive and probably would have never happened,” as Seema points out, due to the sheer scale and complexity. To overcome this, the project leveraged the expertise of the camp’s residents. Humanitarian OpenStreetMap (HOT) partnered with locals to collect imagery and collaboratively label data, recognizing that “no one knows better their community than the people that are living there.” This participatory approach ensured the accuracy and relevance of the map data. Duncan highlights the detail captured, noting the inclusion of not just solar panels but also “those poles,” demonstrating the comprehensive nature of the data collection.

AI-Powered Mapping & Open-Source Tools

A subsection of the mapped area was used as a “training dataset for an AI model.” Simone explains that the “green dots are ground truth,” referring to the manually verified data used to teach the AI to identify features. The AI for Good Lab then developed open-source code capable of mapping various elements including solar panels, buildings, roof types, and sanitation facilities. Amrita emphasizes the efficiency of this approach: “Once you’ve trained a model on a small amount of data it’s very fast to get predictions on new areas.” This allows for rapid mapping of larger areas beyond the initial training dataset.

Synergies Between Domain Expertise and AI

The project’s success stemmed from the combined strengths of different teams. Simone highlights this synergy, stating, “Each team brought something unique to the table. The domain experts understand the human aspect, the implications on the ground, and then we were able to bring in the AI expertise.” This integration ensured that the technological solutions were grounded in real-world needs and considerations.

Open Source & Scalability for Global Impact

A core principle of the project is open-source accessibility. Juan articulates this philosophy: “Our job is to use our knowledge and our technology to help others…That’s the magic of open source.” The open-source nature of the code allows any organization or refugee camp to utilize, adapt, and contribute to the mapping efforts. The ultimate ambition, as stated by Juan, is to “help every refugee camp to be able to use drone imagery to create their own maps, now and in the future,” extending the benefits beyond Kakuma.

Conclusion

This initiative demonstrates the powerful combination of community-based mapping, artificial intelligence, and open-source technology to address critical challenges in humanitarian settings. By empowering local communities and providing accessible tools, the project aims to improve logistical planning, expand access to essential services like electricity, and ultimately foster economic opportunities for refugees. The emphasis on open-source principles ensures scalability and long-term sustainability, paving the way for similar mapping efforts in refugee camps worldwide.