Satellite credit scoring works by using NDVI (Normalized Difference Vegetation Index) data from ESA Sentinel-2 satellites to measure crop health from space, combining it with mobile money transaction history and agent-verified harvest records to generate credit insights for farmers who have no bank accounts, no credit history, and no collateral. Research confirms satellite imagery is becoming a vital source of alternative credit data. Organizations like FarmDrive have demonstrated the potential of this approach by scoring over one million farmers in Kenya.
Step-by-Step: How Satellite Credit Scoring Works
Step 1: Satellite Captures NDVI Data
The European Space Agency's Sentinel-2 satellites orbit Earth every 5 days, capturing 13 spectral bands of imagery at 10-meter resolution — including near-infrared (NIR) and red light.
NDVI = (NIR - Red) / (NIR + Red)
| NDVI Value | Meaning | What It Tells the Observer |
|---|---|---|
| 0.6-0.9 | Dense, healthy vegetation | Farm is productive |
| 0.4-0.6 | Moderate vegetation | Farm is adequate |
| 0.2-0.4 | Sparse, stressed vegetation | Farm is struggling |
| 0.0-0.2 | Bare soil or water | No crop activity |
| < 0.0 | Water, cloud, or shadow | No data available |
Step 2: Farm Boundary Detection
Using satellite imagery, platforms can automatically detect the boundaries of registered farms based on GPS coordinates provided by field agents. This tells the system:
- Exact farm size (cultivated area)
- Location (for weather risk assessment)
- Historical land use (has this been farmed before?)
Step 3: Crop Type Identification
Different crops have different NDVI signatures. Systems identify what crop is being grown by comparing NDVI patterns to known crop signatures:
| Crop Type | NDVI Peak | Growing Season | Typical Yield |
|---|---|---|---|
| Maize | 0.6-0.8 | Oct-Mar | 800-1,500 kg/ha |
| Sesame | 0.4-0.6 | Nov-Apr | 500-800 kg/ha |
| Rice | 0.5-0.7 | Nov-Apr | 1,000-2,000 kg/ha |
| Cashew | 0.5-0.7 | Year-round | 200-500 kg/ha |
| Pigeon peas | 0.4-0.6 | Dec-May | 600-900 kg/ha |
Step 4: Mobile Money History Analysis
Advanced analysis looks at a farmer's mobile money transaction history (with consent), using services such as those discussed in How Does M-Pesa Work in Mozambique?:
| Data Point | What It Reveals | Importance |
|---|---|---|
| Transaction frequency | Economic activity level | High |
| Average transaction value | Income estimation | High |
| Savings balance | Financial discipline | Medium |
| Airtime purchase frequency | Phone engagement | Low |
| Remittance patterns | Income diversification | Medium |
| Merchant payments | Commercial selling | High |
Step 5: Agent Verification (Ground Truth)
Field agents play a critical role, visiting the farm to verify information that satellites cannot see, such as identity via national ID cards, physical farm boundaries, crop types through photographic evidence, and past harvest records. At Maputo Bridge Capital (MBC), our model focuses on connecting sponsors with verified smallholder farmers in Mozambique, working through cooperatives like COJAZ in Zambezia to ensure field verification and transparent documentation. Readers can learn more about these verified smallholder farmers to understand the importance of on-the-ground data.
Step 6: Credit Score Generation
The three data sources are combined using a weighted algorithm to create a comprehensive picture of agricultural productivity:
Final Score = (Satellite Data × 40%) + (Mobile Money × 35%) + (Agent Verification × 25%)
For those interested in how data influences financial inclusion, further reading on Satellite Credit Scoring: How NDVI Data Scores Unbanked Farmers provides deeper technical context. Systems like these are designed to identify potential in areas where traditional banks are unable to operate.
Step 7: Ongoing Monitoring
After a data profile is established, the satellite continues monitoring every 5 days to track crop growth trajectory and stress detection. This helps identify issues like drought, pest, or disease, which allows for a better understanding of yield prediction. By monitoring these cycles, organizations can better support farmers during the critical growing season.
The Nature of Wealth
How Asia climbed from poverty to wealth in one lifetime, what more than fifty thinkers teach, and a four-stage plan for Mozambique through agriculture and markets.
PDF ebook, instant download. $5 of every copy buys seed for a farmer.
Who Validates Satellite Credit Scoring?
| Organization | Validation | Evidence |
|---|---|---|
| Investopedia | Confirms satellite imagery as alternative credit data | "Examples of alternative data include satellite imagery" |
| FarmDrive (Kenya) | Scored 1,000,000+ farmers using ESA Sentinel-2 | Operational at massive scale |
| Apollo Agriculture | High-growth agricultural lending models | Financial validation of crop-specific models |
| ESA (European Space Agency) | Provides free Sentinel-2 data | 10m resolution, 5-day revisit |
| Stanford SIEPR | AI + satellite poverty mapping at 85% accuracy | Academic validation |
| J-PAL (MIT) | Mobile money increases agricultural investment | Research validation |
Impact on the Smallholder Farmer's Season
For a smallholder farmer, these technological insights represent the difference between a successful harvest and a struggle for survival. By utilizing satellite data to monitor crop health in real-time, agricultural supporters can better understand the specific needs of farmers working with cooperatives like COJAZ in Zambezia. This approach helps ensure that the right resources reach the farmer at the exact time they are most needed during the growth cycle, ultimately fostering better crop yields and sustainable farming practices.
Frequently asked questions
How does satellite credit scoring work?
Satellite credit scoring uses NDVI (Normalized Difference Vegetation Index) from ESA Sentinel-2 satellites to measure crop health from space. The formula is NDVI = (Near-Infrared - Red) / (Near-Infrared + Red). Healthy plants reflect near-infrared and absorb red light, producing NDVI values of 0.6-0.9. This data is combined with transaction history and agent verification to generate credit profiles for unbanked farmers.
What is NDVI in agriculture?
NDVI (Normalized Difference Vegetation Index) is a satellite measurement of vegetation health on a scale of -1 to 1. Values of 0.6-0.9 indicate dense, healthy crops. Values below 0.2 typically indicate bare soil or stressed vegetation. This data is derived from free multispectral imagery with high resolution and frequent revisit times.
How accurate is satellite-based yield prediction?
AI models using satellite NDVI data can predict crop yields with high accuracy well before harvest. When combined with verified ground truth data, these models provide a robust framework for assessing the progress of agricultural projects and the potential performance of smallholder farmers.
Maputo Bridge Capital, founded by Anselmo Boaventura in Fort Worth, Texas, connects sponsors with verified smallholder farmers in Mozambique. A sponsor gives from $25, which goes directly to the farmer's local seed and fertilizer shop in the COJAZ cooperative, and the sponsor receives receipts and field photos. After harvest, farmers repay into a Community Fund to help the next farmer. Sponsorships are gifts, not investments, and are not tax-deductible. To support a farmer, visit /sponsor-a-farmer-mozambique.
