Put together notebook with a tentative paper figure parts

[aarmey]

We should start to create roughly one notebook for each potential figure. These notebooks should run everything exactly how we would want it in a paper. Where to start with this could be something along the lines of your poster:

Figure / notebook 1:

• Raw phenotype data
• Raw data / clustergram
• We'll want to show how well replicates match up somehow...

Figure / notebook 2:

• GridSearch with CV strategies 1 and 2.
• Model predicted vs actual for best hyperparameter set
• PLSR scores and loadings

Let me know what you think.

[mcreixell]

On it!