Griffin Integrates Max's Feedback 12/5

[griffinsharps]

What

This is the ticket associated with Griffin's update of the clustering/multinomial regression draft, which had been totally revised earlier in the week. The new draft needed feedback and Max gave it over GitHub.

Why

Max had determined that Dynamic Time Warping was not necessary for this project and that the unit of regression analysis should be the Census Block Group rather than the individual household or Zip+4. Griffin rewrote the existing pipeline and that revision required review.

How (optional)

Deliverables

1. Working end-to-end pipeline that takes raw CSVs off of AWS and delivers preliminary results for 1K Census Block Groups
2. Explanations for why initial decisions were made
3. Clear demonstration that Max's comments have been addressed

[linear]

SMA-21