Spatio-temporal modeling of maize prices in data sparse environments (Africa, Asia)

Description of the topic

The ability to map local prices of inputs (e.g. fertilizer) and outputs (e.g. maize grain) for African smallholders would be of tremendous potential utility. Such predictive maps could be used to target investments, or to refine empirical assessments of farm behavior from survey data. While some price prediction methods exist, much more empirical validation remains to be done for rural areas in sub-Saharan Africa.


To evaluate alternative methods for spatio-temporal price prediction.

Work expectations

Student would work under the guidance of CIMMYT project leader, but would need to be self-directed and be capable of working independently, as well as part of a team. We would expect to produce a paper from this work. Intern would work closely with CIMMYT scientists in the Nairobi office, but may also work in field data collection.


  • Assemble data for analysis
  • Implement and evaluate alternative price prediction models
  • Write up analysis

Required skills

Student would need to have programming, GIS & machine learning skills.