docu_Gab

**problem statement: 
1)
each players has their own strenght and value can't compare a goalkeeper with a stiker.
too much information, some are not necessery.
so we decide to scale it down to players with same position and select just a small amount of attributes.
2)
data irrealistic (some players have 0 app with positive number in the field "played time"),
(or an avg. time of over 120 min with a low appearance),
missing data,
3)
what is essential for analysis the player performance? 
ex: (time played in total/appearances)

models:

**thoughts:
Factors for Predicting Real-Life Performance(?)
When using FIFA stats to predict real-life performance.

consider:
Relevant Features:
Identify which FIFA stats (like pace, dribbling, shooting accuracy) are most predictive of real-life outcomes like goals scored, passes completed, or tackles made.

**BONUS: External Factors:
external factors(that FIFA might not capture) 
such as player injuries, team dynamics, or psychological factors.


**Solution:

1)we want to predict if the fifa overall ability can reflect the real life rating.
>>use these 2 attribute to create an overall look first and then
> devided into 3 categories, OFF, DEF, MID and see how accurate it is.

use scattterplot.



**Some extra steps
- Added column "Avg min / game" in Excel