Schedule August 26 - 30, 2022

Day 1

	Friday August 26, Day 1/3 - Introduction to Crop Condition Analysis
	Activity
1.	Greetings
2.	Current approach to crop condition and yield analysis in RCMRD
	BREAK
3.	VIDEO: How’s it Growing?
4.	VIDEO: AgMet graphics
5.	How to use AgMet graphics?
6.	VIDEO: Using the GLAM system
	LUNCH
7.	Walkthrough of geoprepare library
8.	Walkthrough of key data inputs for geoprepare
9.	Initiate assignment
	BREAK
10.	* Assignment * Check geoprepare installation on RCMRD and NASA servers with Benson and Sara

Assignment

Day 1 assignment

Day 2

	Monday August 29, Day 2/3 - Crop Condition Analysis
	Activity
1.	Review of day 1
2.	Install MOBAXTerm, access RCMRD cluster
	BREAK
3.	Creating a crop condition report
4.	Assignment
	LUNCH
5.	Building a simple crop yield forecast model
6.	Assignment

Assignment

Link to AgMet and percentile plots: https://www.dropbox.com/sh/40343f0zvyidsbm/AADjkiwtEpLY5bAs395w3rTya?dl=0
Document: https://docs.google.com/document/d/1Nz5x-R7Kl4DDxZ0uEe453QrhXwci7K8tpu3ZYGl6G4I/edit?usp=sharing

• Perform crop condition analysis for Kenya, Malawi, Rwanda, Zambia and United Republic of Tanzania using the graphics and questions outlined here

• Repeat all the 4 models from here for each of the following countries:
  Malawi
  Zambia
  Rwanda
  United Republic of Tanzania
  Put all the results from the 4 models and 5 countries in a table and compare them.

• Use a Random Forest model instead of linear regression for all 4 models and 5 countries. What are the results?
  HINT 1:
  A random forest regressor is non-parametric model and does not have any coef_ and intercept_ attributes.

  HINT 2:

  from sklearn.ensemble import RandomForestRegressor  
  model = RandomForestRegressor(n_estimators=250, random_state=0)
  

Which model performs better and under what conditions?