Improving Food Security and Crop Yield in Kenya Through Machine Learning

Challenge Background

Based on data collected during the 2020 short rains assessment, the Kenya Food Security Steering Group (KFSSG) estimates that around 1.4 million Kenyans in arid and semi-arid areas are facing Crisis (IPC Phase 3) or worse outcomes, an increase of 93 percent compared to the preceding long rains season. Cumulatively below-average rainfall across eastern Kenya resulted in a poor harvest in marginal agricultural livelihood zones and declines in rangeland resources in pastoral areas driving Stressed (IPC Phase 2) and Crisis (IPC Phase 3) outcomes across northern and eastern Kenya.

The Problem

More than 1.4 million Kenyans in arid and semi-arid areas are facing food crises or worse outcomes. Covid-19 control measures, desert locust invasion, and climate change have negatively impacted crop production and rangeland resource regeneration.  

With the help of Machine Learning, farmers should be able to predict weather patterns and conditions in different places in Kenya for the next farming season while promoting a data-driven agricultural system. Data such as soil PH, temperature, and moisture levels, land usage, combined with other data sources from World Bank’s data portal and Kenya Meteorological Department could be processed to show exactly when and where farmers should improvise their farming method, and to know the best crop type to be cultivated. The data will also help to decide where to invest, and make use of unutilized land for farming.

Goal of the Project

• Identify un utilised farming land through satellite imaging.
• Applying Kenyan based open-source satellite imagery dataset to make crop yield prediction.
• Create a weather information sharing system for farmers for better farming decisions.

Project Timeline

What you'll learn

1. Satellite Image data collection

2. Weather patterns analysis

3. Computer Vision for crop type detection

4. Data visualization using pandas, matplotlib and QGSI