Overview

Causal analysis, understanding the effect or impact that reviews have on revenue.

Linear regression, we may transform the outcome to the log scale.

Stated Factors
Movie Genre
Production company, 
Movie budget
IMDB user ratings
Studios - are more immune to critic reviews than others? 
Big-budget films & small-budget?

Approach
1. Build a non-linear, tree-based model (i.e., RF or GBDT) to predict Revenue based on all of the stated factors.
2. For each subgroup we are to analyze, predict E[Rev|Good Review, Subgroup] and E[Rev|Bad Review, Subgroup] 
3. Plot the above against each other for different subgroups. Those that are on the diagonal are those that are more immune to bad reviews. 

Drop Column V in movie data with metacritic data.
Drop column R
Drop column F
Drop column E
Check duplicates using column for original movie title.
How do you determine if data is enough data?
Gross or Net revenue
International or Domestic revenue we are assuming worldwide data