Visualization of data extracted from sports - mainly with python and matplotlib
- theme - (go to code)
Clustering teams
we want to group the teams basically according to their xG Points and their Points - for this we use kMeans clustering, previously we determined the number of groups in 6 clusters
- theme - (go to code)
Power Space - based on real performance and calc xG perfomance
The team with the most points is 100% - the value of the others is relative to them. We were looking for a gap between the total value of the teams and 200% (real 100% + xG 100%)
Questions:
- How big is the gap?
- In which direction is greater or balanced?
- theme -
the points scored by the teams in relation to the total points that can be scored in the 5-match grouping
plus: mood change according to loss aversion/prospect theory measured in util - comparative (go to code)
- theme - Random walk with the positions - comparative (go to code)
- theme - Real table vs. xG table - comparative (go to code)
- theme - xG Match flow (go to code)
- theme - position of teams per round - comparative (go to code)