# How Satellites Are Supporting Farmers Across Africa | Catherine Nakalembe | TED

Source: https://www.youtube.com/watch?v=pWKggqO3VlY
Recap page: https://rapidrecap.app/video/pWKggqO3VlY
Generated: 2025-11-20T16:38:35.855+00:00

---
## Quick Overview

Food security specialist Catherine Nakalembe details how utilizing satellite data, AI, and on-the-ground efforts by volunteers and motor-taxi drivers is empowering farmers across Africa, particularly in Uganda and Kenya, to better predict and adapt to threats like drought, pests, and floods, thereby saving livelihoods and improving crop yields.

**Key Points:**
- Satellite data, AI, and on-the-ground data collection via volunteers and motorbike taxi drivers are crucial for supporting smallholder farmers across Africa.
- The complexity of African farming, characterized by small, diverse plots (like the tapestry of rice paddies shown), makes traditional modeling inadequate.
- Catherine Nakalembe focuses her work on several African nations including Uganda, Kenya, Tanzania, Zambia, Mali, and Senegal.
- The collected data is used to map what crops are growing where, predict rainfall, forecast flood impacts (like the 2024 Kenya floods), and estimate total affected crop land.
- In Kenya, a rapid assessment following a flood used this data to inform the Ministry of Agriculture's response programs, which previously lacked accurate damage estimates.
- The methodology involves equipping motorbike taxi drivers (Boda Boda) with GPS-enabled GoPros to capture ground imagery, which is then cross-referenced with satellite data to train better AI models.
- The ultimate goal is to create more complex, context-aware models that move beyond generalized Western agriculture data to provide actionable, localized insights for subsistence farmers.

![Screenshot at 0:17: An aerial shot showing arid, dry farmland surrounding a small cluster of rural homes, illustrating the challenging environmental context where many African smallholder farmers operate.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-00-17.png)

**Context:** Catherine Nakalembe, a Food Security Specialist and TED Fellow, discusses the challenges faced by smallholder farmers in Africa, where traditional farming methods struggle against increasing climate variability, pests, and diseases, leading to devastating crop failures. Her work centers on leveraging modern technology, specifically satellite imagery and AI, combined with grassroots data collection, to create localized agricultural intelligence that directly supports vulnerable farming communities.

## Detailed Analysis

Catherine Nakalembe explains that the challenges faced by smallholder farmers—such as drought, pests, and floods—are demoralizing when there is no recourse, especially when farming is their sole source of income. She highlights that agricultural models trained on data from large, single-crop farms in Europe or the US often fail to accurately represent the complex, diverse, and small-plot farming systems common across Africa. To address this, her team uses remote sensing data from over 8,000 satellites observing Earth daily, combined with AI processing. Crucially, they bridge the gap between satellite data and ground truth by employing local volunteers and motorcycle taxi drivers (Boda Boda) equipped with simple GoPros to capture street-level imagery of fields and infrastructure. This ground data is used to train AI models to accurately map crop types and monitor conditions like flood impact or crop health across specific regions in countries like Kenya and Uganda. For instance, after rapid flooding in Kenya in 2024, this data was used to provide the Ministry of Agriculture with an accurate estimate of affected farmland, something impossible with outdated methods, allowing for targeted relief programs. The ultimate aim is to develop sophisticated models that learn from diverse local contexts to save livelihoods and provide timely, actionable information to farmers.

### The Problem

- Farmer Vulnerability
- Farmers face demoralizing outcomes from drought, pests, and floods when farming is their only income source
- Traditional models fail because African farms are small and diverse, unlike large-scale Western agriculture.

### The Solution

- Data Integration
- Utilizing data from over 8,000 satellites observing Earth daily, processed with AI
- Bridging the gap by collecting ground truth data using local volunteers and Boda Boda drivers with GoPros.

### Field Operations

- Data Collection
- Motorbike taxi drivers wear GoPros while driving through farms in Uganda and Kenya to capture imagery
- This data is used to train AI models to recognize specific crops (maize, beans, cassava) and environmental conditions.

### Case Study

- Flood Response in Kenya
- Following the 2024 floods, satellite data provided an estimate of total affected crop land
- This information informed the Ministry of Agriculture's response programs, preventing the use of inaccurate data.

### The Goal

- Context-Aware Modeling
- The objective is to move beyond generalized models to create complex systems that learn from diverse local contexts
- This allows for providing highly specific and actionable information to support farmers.

![Screenshot at 0:04: Aerial view of a small, arid village surrounded by sparsely cultivated brown fields, illustrating the difficult agricultural context.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-00-04.png)
![Screenshot at 0:06: Close-up of withered, dry corn stalks against a backdrop of a large sand dune, symbolizing crop failure due to environmental stress.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-00-06.png)
![Screenshot at 0:12: Catherine Nakalembe, Food security specialist and TED Fellow, speaking in an interview setting.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-00-12.png)
![Screenshot at 0:43: A laptop displaying Google Earth Pro, showing remote sensing data overlaid on a map, demonstrating the technological tools used for spatial analysis.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-00-43.png)
![Screenshot at 3:55: A satellite map displaying large blue areas overlaying a river system \(Tana River County\), visualizing the extent of flood inundation on agricultural land.](https://ss.rapidrecap.app/screens/pWKggqO3VlY/00-03-55.png)