Mining Video Game Data

Most recently, I decided to do a more in-depth comparative analysis of several Super Nintendo role playing games based on game data, to determine whether there is a link between specifics of the game data and the number of sales of the game in various countries/regions, where the sales data is available. In particular, these older SNES role playing games have a variety of data that can be analyzed: graphics data (sprites, etc); music data; and text data.

My study of this information is to determine whether or not certain characteristics seem to provoke greater sales of the game. It would be interesting if games utilizing certain colors in their sprites more often, or certain musical notes or chords, or certain text themes might be linked to increased sales/consumer purchasing behavior. Then again, perhaps such a study would simply show that drawing such conclusions are not possible. Either way, the goal is to explore this and determine some facts about this.

Currently, the decision is to analyze data from the following video games:
Final Fantasy Mystic Quest - developed by Square
Chrono Trigger - developed by Square
Secret of Mana - developed by Square
Earthbound - Ape Studios / HAL Laboratory
Breath of Fire II - developed by Capcom
Illusion of Gaia - developed by Quintet
Lufia & the Fortress of Doom - developed by Neverland
Ys III: Wanderers from Ys - developed by Nihon Falcom

The current plan is to scrape data from certain websites using Python, or manually where needed-- the websites used will be noted in the analysis. I plan to use other software at my disposal (R, SPSS Modeler,etc), and will share code/results as this progresses. My goal is mainly to have fun and learn some new things, and hopefully some interesting knowledge will be squeezed out along the way.