This Python script collects and analyzes shooting data for a football team and its opponents from FBref for multiple seasons. The primary goals are to calculate rolling averages for expected goals (xG) and expected goals against (xGA), and then visualize the trends over time.

• pandas
• time
• matplotlib.pyplot
• numpy

The script iterates over specified seasons, constructs URLs for match logs on FBref, and retrieves shooting data for the football team and its opponents. The data is processed, including renaming columns and merging based on common identifiers.

This is where you can find the team_id in fbref:

Rolling averages for xG and xGA are calculated using a window size of 19 matches. The values are rounded to two decimal places for clarity.

df['xG_rolling'] = df['npxG_for'].rolling(window=19).mean()
df['xGA_rolling'] = df['npxG_against'].rolling(window=19).mean()
df = df.round(2)

The script utilizes matplotlib to create a plot with a black background. Rolling averages for xG and xGA are plotted using distinct colors. Season dividers are represented by grid lines with a specific color. Connecting lines between xG and xGA for each match provide additional insights.

plt.plot(df.Match, df.xG_rolling, color=xg_clr)
plt.plot(df.Match, df.xGA_rolling, color=xga_clr)

for match, xg, xga in zip(df.Match, df.xG_rolling, df.xGA_rolling):
    plt.plot([match, match], [xg, xga], color=connector_clr, linestyle='--', linewidth=2, alpha=.5, zorder=1)

The resulting plot is saved as an image file in the 'Output' directory with a filename based on the team name and the metric ('-RollingXg.png').

plt.savefig('Output/' + team + '-RollingXg.png', bbox_inches='tight')

This has a runtime of about a minute for five seasons.

For the player edition, the rolling avg is for 10 games.