Skip to content

FerSotoApse/Score_methods

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Score Methods Study

Streamlit app link

Resume

Common scoring systems used in games consider a simple accumulation of points gained by the user, and tend to gain complexity depending on what is being measured. This methods work well when are applied on single players or on team games with a defined, unmutable size, but not in a situation were the size is defined by user preferences, were teams or categories will be naturally unbalanced. Here is proposed an alternative scoring method based on the performance of each group, which can be applied on large, naturally irregular groups, tested on a event escenario with multiple activities.

Data description

All initial synthetic data is randomly generated, and based on the user point of view interface:

  • player_id: user unique identification
  • event_date: date of the event, in YYYY-MM-DD format
  • event_game: activity within the overall event, can be multiple activities happening at the same time
  • team: team, group or category where the user belongs.
  • score: user score in an activity. 0 is given if not participated in the activity, 1 to 3 if the user participated
  • medal: medal won by the user in an activity. 0 is 'not played', 1 is 'bronze', 2 is 'silver', 3 is 'gold'

Method

Sources

Libraries

Data generation and manipulation

  • Pandas
  • Numpy
  • Random
  • Datetime

Machine Learning Model

  • Scikit-learn for KMeans unsupervised model and metrics

Data visualization

  • Plotly graph_objects
  • Plotly Express