Python is a programming language. For many years I have used python for data analysis; the NCAA 14 dynasty save is basically a big database (think of a bunch of excel sheets filled with data). So I export tables from the EA DB editor, run my python scripts to perform edits to them, then re-import them with the EA DB Editor.
For instance, I export the "PLAY" table from my dynasty save which contains all of the player attributes. My python script modifies every player's attributes based on a random number, and based on the hidden "discipline" attribute. The mean of my added progression is -2% each season; so on average, a redshirted player with 4 years of progression will be 8% lower in overall than he would have been with the default in-game progression. So a player that would have been a 90 OVR before would be an 83 on average. No more building a monster out of unrecruited 3* players.