phil_exploration.R

## 1a) How do ad groups perform against each other?

## 1b) How do campaigns perform against each other?

## (Beware of event_id because of the many-to-many
## relationship in GA and Purchase.)

## 2) How do customers behave (on different variables)
## in each event category? 

## 3a) Who are the true fans who 1. buy tickets only
## from TM and 2. always go to the events themselves
## and don't resell them. Who are they? Are there
## distinct characteristics in each cluster?

## 3b) Who are the scalpers? hey buy tickets early (What
## is early?) and sell them later at a much higher 
## price on the resale market.

## 4a) Discover other segments among our user base 
## (e.g. one time buyers)?

## 4b) How would one advertise/market to these segments?
## Correlations? Causations?

## 5 How to optimize pricing for each location or 
## different types of event and different people?

## 6 How to generate keywords for small or infrequent
## shows and price them? Currently the Adgroup
## "ZZ - Bob's Automation Shop" does it. How
## does it work in terms of the ads generated?

## 7 How to do word association so we get associated
## keywords like "Orange" for SU and "Blue Devils"
## for "Duke" (idea: use packages "twitterR", "tm", 
## "ggplot")