ac209_finalproj_notes.txt

AC209 
Final Project Notes
A. Meisner & M. Williams

Meeting Notes
_______________
_______________



10.24.2014
__________
** Possible Questions/Directions
	- Predict the positve/negative score of a game based on the text of the associated review 
	- Develop a regression model to predict the numerical score of a game given text in the review 
	- Use reddit or other social networking sites plus gaming sites to predict scores for future games 

** Possible Data Sets to use
	(Gaming)
	- Giant Bomb (has an API): http://www.giantbomb.com/api/
	- IGN
	- Gamespot
	(Social Sites)
	- Twitter
	- Reddit

** Main Tasks for next meeting
	- Get Giant Bomb API and try to get any form of data
		- If Fail: Make preparations to Data Scrape 





11.08.2014
__________
** Results since previous meeting
	- Success in accessing Giant Bomb data through API

** Main Question 1: Can we predict the score of a review based on the text in the review blurb?

** Main Question 2: (Need additional questions for analysis)

** Exploratory Analysis Questions (look at "gp_expl_analysis.ipynb" for details)
	- Giant Bomb score distribution
	- Correlation between game score and company
	- Types of games which are the most popular
	- Giant Bomb scores over the years
	- User score distribution
	- User scores and critics (i.e. staff) scores
	- Staff score distribution



** Main Tasks for next meeting
	- Scrape reviews from Game Spot if possible
	- Rough Draft of Proposal?
	- Complete Exploratory Analyses on Giant Bomb Data Set

	- Create a stable data set: 
		- Review information 
		- Company information
		- User Reviews