internship_club_brugge

A summary of the notebooks created during the internships The headers represent the folders containing all work per project The bold subheaders represent specific notebooks

Inwerkdag

This directory contains data and notebooks of my first few days at the internship. Where I focused on getting familiar with python

Competition Comparison

exploratory_analysis.ipynb
Contianing code of the initial data exploration
Investigates correlations between metrics

benchmarking.ipynb

• Combines Skillcorner match data with statsbomb competition information and in house quality scores
• Normalizes data to 90 minutes
• Calculates statistical metrics for for physical match data per competition and position
• Writes the statistical data to 2 seperate excel formats
• Stores processed data in a single parquet file

position_benchmarking_script.py
Contains the same content as _ benchmarking.ipynb _ but is a single script

statistical_testing_competitions.ipynb
Identifying competitions that have similar demands

• Compares competitions based on a selected match metrics
• Attempts to find statistically significant differences using Tukey HSD post hoc analysis
• Results are written to 2 Excel files, one file compares the competitions on a general level, the second file compares the competitions per position.

clustering_competitions.ipynb
Identifying competitions that have similar demands

• Clusters created based on 2 metrics using K-means from Scikit-Learn
• Optimal cluster number is detrmined using silhoutte score (limit set to 6)
• Results are written to an Excel file

visualisation.ipynb
Notebook containing Dash app

• Compares competion and position demands
• Clusters competitions

Positional data

The aim of this directory was to process Secondspectrum optical tracking data from matches using various methods

floodlight_functions.ipynb
Contains modified/optimized versions of the functions of the Python Floodlight module

floodlight2parquet.ipynb
Extracts meeaningfull data from the Secondspectrum json files using Floodlight functions

• X-Y coordinates per frame dataframe
• Distance moved per frame dataframe
• Moving velocity per frame dataframe
• Using slight changes acceleration dataframe can also be determined
• Dataframes are written to seperate parquet files to avoid unescessary repetitio, of the heavy Floodlight functions

NumberOfRuns.ipynb
Calculates the total distance ran above or between certain speed tresholds per player using the Floodlight distance and velocity dataframes

velocity_distance.ipynb
Calculates the number of runs above or between certain speed tresholds per player using the Floodlight velocity dataframe

ChangesOfDirection.ipynb
Calculates the number of direction changes per player

• Contains 3 different calculation methods
• Reilly, B., Morgan, O., Czanner, G., & Robinson, M. A. (2021). Automated Classification of Changes of Direction in Soccer Using Inertial Measurement Units. Sensors, 21(14), 4625. https://doi.org/10.3390/s21144625
• Merks, B., Frencken, W., Otter, A. R., & Brink, M. (2021). Quantifying change of direction load using positional data from small-sided games in soccer. Science and Medicine in Football, 6. https://doi.org/10.1080/24733938.2021.1912382
• Kai, T., Hirai, S., Anbe, Y., & Takai, Y. (2021). A new approach to quantify angles and time of changes-of-direction during soccer matches. PLoS One, 16(5), e0251292. https://doi.org/10.1371/journal.pone.0251292

ChangesOfDirection_PossessionStatus.ipynb
Calculates the number of direction changes per player both in and out ball possession