The Feed the Future Myanmar Agriculture Policy Support Activity (MAPSA) is hosting a series of capacity strengthening modules designed to train Myanmar economists and policy analysts on current knowledge, evidence, and methods for the analysis of the Myanmar agri-food system. Through these modules, participants will learn new analytical and relevant economic methods using updated datasets for Myanmar and will benefit from expert insights on the functioning of the rural and agricultural economy. Through this series, participants will gain a comprehensive understanding of the supply, demand, and markets for agricultural products, nutritional issues, the non-farm economy, and investments required to achieve inclusive and sustainable growth in Myanmar.
About the Module
Monitoring the agrifood system in fragile states, such as Myanmar, is challenging. The ability to collect data in these environments can often be costly and access to respondents constrained due to conflict and other restrictions. This limits the guidance for spatially targeted interventions to reach the most vulnerable populations. This webinar will showcase three innovative uses of satellite imagery and machine learning to overcome these challenges and monitor the state of agriculture and food security in Myanmar.
Small-scale Food Security Mapping in Fragile and Conflict-affected Countries: A Machine Learning Approach
Zhe Guo, Senior GIS Coordinator, International Food Policy Research Institute
This presentation will demonstrate the use of data-driven approaches such as machine learning in combination with high-resolution satellite imagery and geospatial datasets to fill knowledge gaps about poverty and welfare indicators in geographical regions with limited access. Through this approach and based on high-frequency survey data collected by the Myanmar Agricultural Policy Support Activity (MAPSA), we can generate timely predictions on welfare indicators in Myanmar with high spatial resolution to guide effective and targeted resource allocation and policy actions. The data-driven models also pinpoint key parameters affecting welfare levels, including past-month precipitation, altitude, proximity to the neighboring country’s border, and nightlight intensity.
Counting Chickens from the Sky
Xiaobo Zhang, Senior Research Fellow (International Food Policy Research Institute) & Chair Professor (Peking University)
This presentation will track the performance of integrated poultry-fish production in Myanmar by combining information from pre-existing field surveys with remotely sensed data. We frame the problem as a deep learning task, including object detection, object segmentation, and cluster detection. Our trained model identified 1,750 integrated poultry-fish farms in Yangon region based on satellite imagery in 2018, closely matching the number of farms reported in surveys conducted around this time. Our model reveals that the number of poultry farms in Yangon contracted sharply from 2020-2023, reflecting the combined impacts of COVID-19 restrictions and the political instability.
Rice Mapping in Myanmar 2017-2023
Sarah Kanee, Remote Sensing Data Analyst, Asian Disaster Preparedness Center
This presentation will provide an overview of rice mapping in Myanmar utilizing satellite imagery and applying machine learning classifications to generate maps and derive rice area from image-interpreted samples. Using available national statistical data, we can estimate yields and production of the identified areas under cultivation. An overview of the rice estimates per region across the years 2017 – 2023, along with an update on the work for the Monsoon 2023 season, will be presented.
Moderator
Ian Masias, Deputy Chief of Party, Myanmar Agriculture Policy Support Activity, International Food Policy Research Institute
Download the slides from "Insights from Above: Satellite and Machine Learning for Agriculture and Food Security in Myanmar " here.
