From Pixels to Plots: Transforming Lives of African Smallholder Farmers through Geospatial Data
summary
One Acre Fund applies Python pipelines and Google Earth Engine to serve about 4 million farmers (hit 1M in 2019) and measures impact with an SROI up to 7.5 in Rwanda and roughly $3.5 per donated dollar overall. The team trains a ResUNet segmentation model with GPU support to delineate plots (optimizing MCC and IoU), finds outputs are roughly 80% useful but miss very small fields, builds a Streamlit app + APIs to serve recommendations, and uses USSD for feature phones; tree tagging works best with a four‑point consensus.
AI-written summary · about
chapters (11)
- 0:00Introduction to the talk
- 0:46About One Acre Fund and its mission
- 3:18Farmer prosperity stages explained
- 5:06Measuring impact on farmers
- 7:35Challenges in locating farmer plots
- 10:55Demo of field delineation tool
- 13:29Tree detection and GPS accuracy issues
- 15:13Using Google Earth Engine for data
- 19:55Deploying recommendations to farmers
- 21:19Technology for farmer communication
- 22:48Q&A session
related talks
01h 36mFrom Maps to Models: An End-to-End Journey in Geospatial Machine Learning with Python
27mThe View from Above Getting Started with Satellite Images
33mBuilding Data Science Tools for Sustainable Transformation
30mUnlocking Geospatial Data Science: A Strategic Overview
46mData Discussion - AI for Good
22mExpert Panel | Location Data in Retail & Real Estate: Delivering an ROI Through Analytics
25mEverything but maps with spatial tools
58mImproving disaster response with machine learning