Tutorial for 4 Nations Face-Off
Intro
Use the chickenstats library to scrape play-by-play data and international team performance in the 4 Nations Face-Off.
Parts of this tutorial are optional and will be clearly marked as such. For help, or any questions, please don't hesitate to reach out to chicken@chickenandstats.com or @chickenandstats.com on Blue Sky.
Housekeeping
Import dependencies
Import the dependencies we'll need for the guide
import datetime as dt
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from matplotlib.offsetbox import OffsetImage, AnnotationBbox
import chickenstats.utilities
from chickenstats.utilities import ChickenSession
from chickenstats.chicken_nhl import Scraper, Season
from chickenstats.chicken_nhl.info import INTERNATIONAL_COLORS, Team
from chickenstats.chicken_nhl.helpers import charts_directory
from pathlib import Path
from PIL import Image
from io import BytesIO
from bokeh.plotting import figure, output_file, output_notebook, show, save
from bokeh.models import HoverTool, ColumnDataSource, Title, Div, Span
from bokeh.layouts import gridplot, column
from bokeh.models import Range1d
Pandas options
Set different pandas options. This cell is optional
Folder structure
Chickenstats matplotlib style
chickenstats.utilities includes a custom style package - this activates it. This cell is also optional
Scrape data
Schedule
Scrape the schedule using the Season object
Game IDs
Create a list of game IDs to scrape
conds = schedule.game_state.isin(["OFF", "FINAL"])
game_ids = schedule.loc[conds].game_id.unique().tolist()
latest_date = schedule.loc[conds].game_date.max()
Play-by-play
Scrape play-by-play data using the Scraper object
Stats
Aggregate statistics to season and game level
Matplotlib
strength_state = "5v5"
toi_min = 5
# Setting filter conditions and filtering data
conds = np.logical_and(lines.strength_state == strength_state, lines.toi >= toi_min)
plot_lines = lines.loc[conds].sort_values(by="xgf_percent", ascending=False).reset_index(drop=True)
# Setting overall figures
fig, axes = plt.subplots(nrows=2, ncols=2, dpi=650, figsize=(8, 6))
fig.tight_layout(pad=2.5)
axes = axes.reshape(-1)
# Getting the averages and drawing the average lines
xga_mean = plot_lines.xga_p60.mean()
xgf_mean = plot_lines.xgf_p60.mean()
# Setting the size norm so bubbles are consistent across figures
size_norm = (plot_lines.toi.min(), plot_lines.toi.max())
# Getting the teams and standings data to iterate through
teams = ["CAN", "FIN", "SWE", "USA"]
team_names = ["CANADA", "FINLAND", "SWEDEN", "UNITED STATES"]
team_names = dict(zip(teams, team_names, strict=False))
# Iterating through the standings data
for _idx, team in enumerate(teams):
# Setting the team # Setting the axis
ax = axes[_idx]
team_info = Team(team)
team_name = team_info.team_name
colors = team_info.colors
team_stats_plot = team_stats.loc[
np.logical_and(team_stats.strength_state == strength_state, team_stats.team == team)
]
# Average lines
ax.axvline(x=xga_mean, zorder=-1, alpha=0.5)
ax.axhline(y=xgf_mean, zorder=-1, alpha=0.5)
# Filtering data and plotting the non-selected teams first
conds = plot_lines.team != team
plot_data = plot_lines.loc[conds]
# They all get gray colors
facecolor = colors["MISS"]
edgecolor = "white" # colors["MISS"]
# Plotting the non-selected teams' data
sns.scatterplot(
data=plot_data,
x="xga_p60",
y="xgf_p60",
size="toi",
sizes=(20, 150),
size_norm=size_norm,
lw=1.5,
facecolor=facecolor,
edgecolor=edgecolor,
alpha=0.5,
legend=False,
ax=ax,
)
# Filtering and plotting the selected team's data
conds = plot_lines.team == team
plot_data = plot_lines.loc[conds]
# Setting the colors
facecolor = colors["GOAL"]
edgecolor = colors["SHOT"]
# Plotting the selected team's data
sns.scatterplot(
data=plot_data,
x="xga_p60",
y="xgf_p60",
size="toi",
sizes=(20, 150),
size_norm=size_norm,
lw=1.5,
facecolor=facecolor,
edgecolor=edgecolor,
alpha=0.8,
legend=False,
ax=ax,
)
# Setting x and y axes labels
x_labels = [2, 3]
if _idx in x_labels:
ax.axes.set_xlabel("xGA per 60 minutes", fontsize=7)
else:
ax.axes.set_xlabel("")
y_labels = [0, 2]
if _idx in y_labels:
ax.axes.set_ylabel("xGF per 60 minutes", fontsize=7)
else:
ax.axes.set_ylabel("")
# Setting tick params font size
ax.tick_params(axis="both", which="major", labelsize=7)
gf = team_stats_plot.gf.iloc[0]
xgf = round(team_stats_plot.xgf.iloc[0], 2)
ga = team_stats_plot.ga.iloc[0]
xga = round(team_stats_plot.xga.iloc[0], 2)
toi = round(team_stats_plot.toi.iloc[0], 2)
# Setting the ax title
ax_title = f"{team_name} | {gf} GF ({xgf} xGF) - {ga} GA ({xga} xGA) | {toi} TOI"
ax.set_title(ax_title, fontsize=7, x=-0.085, y=1.03, horizontalalignment="left")
logo = team_info.logo
imagebox = OffsetImage(logo, zoom=0.15, alpha=0.15, zorder=-10)
imagebox.image.axes = ax
ab = AnnotationBbox(imagebox, (0.5, 0.5), xycoords="axes fraction", bboxprops={"lw": 0}, alpha=0, zorder=-10)
ax.add_artist(ab)
# Figure suptitle and subtitle
fig_suptitle = "Forward line combinations' chances created vs. chances allowed"
fig.suptitle(fig_suptitle, x=0.01, y=1.085, fontsize=11, fontweight="bold", horizontalalignment="left")
subtitle = f"4 Nations Face-Off | >{toi_min} min. TOI at 5v5 (size indicates TOI) | {latest_date}"
fig.text(s=subtitle, x=0.01, y=1.0315, fontsize=10, horizontalalignment="left")
# Attribution
attribution = "Data & xG model @chickenandstats.com | Viz @chickenandstats.com"
fig.text(s=attribution, x=0.99, y=-0.03, fontsize=8, horizontalalignment="right", style="italic")
savepath = Path("./charts/5v5_4n_fo_forwards.png")
fig.savefig(savepath, transparent=False, bbox_inches="tight")
Bokeh
team = "USA"
team_name = "UNITED STATES"
strength_state = "5v5"
device_type = "desktop"
data = lines.copy()
toi_min = 5
display_plot = False
output_file_str = f"{team.lower()}_5v5_{device_type}.html"
plot_attributes = {
"desktop": {
"size_multiplier": 15,
"major_label_font_size": "9pt",
"axis_label_font_size": "10pt",
"title_font_size": "18pt",
"subtitle_font_size": "14pt",
"stats_font_size": "10pt",
"stats_y_value": 0.95,
"player_name_font_size": "10pt",
"ncols": 4,
"height": 240, # 240,
},
"mobile": {
"size_multiplier": 8,
"major_label_font_size": "7pt",
"axis_label_font_size": "8pt",
"title_font_size": "14pt",
"subtitle_font_size": "11pt",
"stats_font_size": "8pt",
"stats_y_value": 0.95,
"player_name_font_size": "10pt",
"ncols": 1,
"height": 240, # 240,
},
}
plot_attributes = plot_attributes[device_type]
plot_colors = {"dark_gray": "#696969", "light_gray": "#D3D3D3", "medium_gray": "#808080"}
# Setting filter conditions and filtering data
conditions = np.logical_and(lines.strength_state == strength_state, lines.toi >= toi_min)
plot_lines = lines.loc[conditions].sort_values(by="xgf_percent", ascending=False).reset_index(drop=True)
x_range = Range1d(plot_lines.xga_p60.min() - 1, plot_lines.xga_p60.max() + 1)
y_range = Range1d(plot_lines.xgf_p60.min() - 1, plot_lines.xgf_p60.max() + 1)
# Getting the averages and drawing the average lines
xga_mean = plot_lines.xga_p60.mean()
xgf_mean = plot_lines.xgf_p60.mean()
# Setting the size norm so bubbles are consistent across figures
size_norm = (plot_lines.toi.min(), plot_lines.toi.max())
# Getting the teams and standings data to iterate through
teams = ["CAN", "FIN", "SWE", "USA"]
team_names = ["CANADA", "FINLAND", "SWEDEN", "UNITED STATES"]
team_names = dict(zip(teams, team_names, strict=False))
plot_list = []
if display_plot:
output_notebook()
else:
output_file(output_file_str)
# Iterating through the standings data
for _idx, team in enumerate(teams):
team_name = team_names[team]
team_stats_plot = team_stats.loc[
np.logical_and(team_stats.strength_state == strength_state, team_stats.team == team)
]
colors = INTERNATIONAL_COLORS[team]
condition = plot_lines.team == team
plot_data = plot_lines.loc[condition].copy()
# plot = prep_stats(data = plot, strengths = strengths)
plot_data["edgecolors"] = colors["SHOT"]
plot_data["colors"] = colors["GOAL"]
plot_data["toi_size"] = plot_data.toi
source = plot_data.copy()
source = ColumnDataSource(source)
TOOLS = "hover,crosshair,pan,wheel_zoom,zoom_in,zoom_out,box_zoom,undo,redo,reset,tap,box_select,poly_select,lasso_select"
p = figure(tools=TOOLS, x_range=x_range, y_range=y_range)
vline = Span(
location=xga_mean, dimension="height", line_color=plot_colors["light_gray"], line_width=2, level="underlay"
)
hline = Span(
location=xgf_mean, dimension="width", line_color=plot_colors["light_gray"], line_width=2, level="underlay"
)
p.add_layout(vline)
p.add_layout(hline)
test = p.scatter(
x="xga_p60",
y="xgf_p60",
size="toi",
source=source,
color="colors",
alpha=0.65,
line_color="edgecolors",
line_width=2.5,
)
condition = plot_lines.team != team
plot_data = plot_lines.loc[condition].copy()
# plot = prep_stats(data = plot, strengths = strengths)
plot_data["edgecolors"] = "white"
plot_data["colors"] = colors["MISS"]
plot_data["toi_size"] = plot_data.toi
source = plot_data.copy()
source = ColumnDataSource(source)
p.scatter(
x="xga_p60",
y="xgf_p60",
size="toi",
source=source,
color="colors",
alpha=0.65,
line_color="edgecolors",
line_width=2.5,
level="underlay",
)
gf = team_stats_plot.gf.iloc[0]
xgf = round(team_stats_plot.xgf.iloc[0], 2)
ga = team_stats_plot.ga.iloc[0]
xga = round(team_stats_plot.xga.iloc[0], 2)
toi = round(team_stats_plot.toi.iloc[0], 2)
# Setting the ax title
subtitle = f"{team_name} | {gf} GF ({xgf} xGF) - {ga} GA ({xga} xGA) | {toi} TOI"
p.add_layout(
Title(
text=subtitle,
text_font_size=plot_attributes["player_name_font_size"],
text_font_style="bold",
text_color=plot_colors["dark_gray"],
vertical_align="middle",
align="left",
text_line_height=1,
),
"above",
)
p.outline_line_width = 0
p.grid.grid_line_color = None
p.axis.minor_tick_line_color = None
p.axis.major_tick_line_color = plot_colors["dark_gray"]
p.axis.major_tick_out = 3
p.axis.major_tick_line_width = 2.5
# p.axis.ticker = [0, 0.25, 0.5, 0.75, 1]
# p.axis.major_label_overrides = {x: f"{round(x * 100)}" for x in [0, 0.25, 0.5, 0.75, 1]}
p.axis.major_label_text_font_size = plot_attributes["major_label_font_size"]
p.axis.axis_label_text_font_size = plot_attributes["axis_label_font_size"]
p.axis.axis_line_color = plot_colors["dark_gray"]
p.axis.axis_line_width = 2
p.yaxis.axis_label = "xGF per 60 minutes"
p.xaxis.axis_label = "xGA per 60 minutes"
hover = p.select(dict(type=HoverTool))
hover.tooltips = [
("LINE", "@forwards"),
("TEAM", "@team"),
("TOI", "@toi{0.0}"),
("GOALS", "@gf"),
("xGF%", "@xgf_percent{0.0%}"),
("xGF", "@xgf{0.0}"),
("xGA", "@xga{0.0}"),
("xGF/60", "@xgf_p60{0.0}"),
("xGA/60", "@xga_p60{0.0}"),
]
hover.mode = "mouse"
hover.name = "tooltips"
hover.renderers = [test]
plot_list.append(p)
grid = gridplot(
plot_list,
ncols=2,
toolbar_location="right",
height=300, # plot_attributes["height"],
sizing_mode="stretch_width",
)
title_str = "Forward line combinations' chances created vs. chances allowed"
title = Div(
text=title_str,
styles={"font-size": plot_attributes["title_font_size"], "color": plot_colors["dark_gray"], "font-weight": "bold"},
margin=(0, 0, 0, 10),
)
todays_date = dt.datetime.now().strftime("%Y-%m-%d")
subtitle_str = f"4 Nations Face-Off | >{toi_min} min. TOI at 5v5 (size indicates TOI) | {todays_date}"
subtitle = Div(
text=subtitle_str,
styles={
"font-size": plot_attributes["subtitle_font_size"],
"color": plot_colors["medium_gray"],
"font-weight": "light",
},
margin=(-5, 0, 10, 13),
background="white",
)
attribution = "Data & xG model @chickenandstats.com | Viz @chickenandstats.com"
footer = Div(
text=attribution, styles={"font-size": "10pt", "color": plot_colors["medium_gray"], "font-weight": "light"}
)
# margin=(0, 0, 0, 0))
chart = column(title, subtitle, grid, footer, sizing_mode="scale_width", background="white")
if display_plot is True:
show(chart, background_fill_color="white")
else:
save(chart)