Make the graphs look cool. It more fun that way.
Data mining is the step that try to take all the raw data collected and turn it into something that makes it possible to make informed decisions about the future of the game. Here is a list of common ones.
How many new users have started the game on a given day. Plot it on a timeline and see if your ads make more people buy the game.
How many unique users played the game over a given period relative to the selected day. DAU is daily users, WAU is weekly and MAU is a whole month. So if you pick a day two months ago it DAU is how many people played the game in the week before the selected day.
The amount of returning players in a Day (D1 or DD), Week (D7 or DW) or Month (D28 or DM). Each value is a % of the player that was new. So D1 shows how many % of the new players played the next day and D7 how many of them played after a week. It does not matter how much they have played in between, just that they play on the given day.