A Python library for plotting hockey rinks with Matplotlib.
## Installation
```pip install hockey-rink```
## Current Rinks
The following rinks are available for use:
- Rink
- NHLRink
- NWHLRink (from the 2021 Lake Placid games)
- IIHFRink
- OldIIHFRink
## Customization
There is also room for customization. For example, to change the dimension of a rink, update the length and/or width of the boards:
```
rink = Rink(boards={"length": 150, "width": 150, "radius": 75})
```
Each rink comes with a default set of features, but additional features can be added. Custom features should inherit
from RinkFeature and override the _get_centered_xy method. The draw method can also be overridden if the desired feature can't be drawn
with a Matplotlib Polygon, though _get_centered_xy should still provide the feature's boundaries. The RinkImage
feature provides an example of this by inheriting from RinkRectangle.
If a custom feature is to be constrained to only display within the rink, the returned object needs to have a
set_clip_path method.
## Plots
There are currently wrappers available for the following Matplotlib plotting methods:
- plot
- scatter
- arrow
- hexbin
- pcolormesh (heatmap in hockey-rink)
- contour
- contourf
- text
There's also a rink.plot_fn which will take as its first argument a plotting method from either Matplotlib or seaborn
and will attempt to make the desired plot.
```
from hockey_rink import NHLRink, RinkImage
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
shots = (
pd.read_parquet("https://github.com/sportsdataverse/fastRhockey-data/blob/main/nhl/pbp/parquet/play_by_play_2023.parquet?raw=true")
.query("event_type in ('GOAL', 'SHOT', 'MISS')")
)
team_colors = {"San Jose Sharks": (0, 0.5, 0.5), "Nashville Predators": (1, 0.7, 0.1)}
rink = NHLRink(
sharks_logo={
"feature_class": RinkImage,
"image_path": "https://upload.wikimedia.org/wikipedia/en/thumb/3/37/SanJoseSharksLogo.svg/330px-SanJoseSharksLogo.svg.png",
"x": 55, "length": 50, "width": 42,
"zorder": 15, "alpha": 0.5,
},
preds_logo={
"feature_class": RinkImage,
"image_path": "https://upload.wikimedia.org/wikipedia/en/thumb/9/9c/Nashville_Predators_Logo_%282011%29.svg/330px-Nashville_Predators_Logo_%282011%29.svg.png",
"x": -55, "length": 50, "width": 29,
"zorder": 15, "alpha": 0.5,
}
)
first_period = shots.query("game_id == 2022020001 and period == 1")
fig, axs = plt.subplots(1, 2, figsize=(18, 8))
rink.scatter("x", "y", facecolor=first_period.event_team.map(team_colors), s=100, edgecolor="white", data=first_period, ax=axs[0])
rink.plot_fn(sns.scatterplot, x="x", y="y", hue="event_team", s=100, legend=False, data=first_period, ax=axs[1], palette=team_colors);
```
When using plots that require binning, it's often best to include a plot_range even when it isn't being used to
find the bins. Here's an example using shooting percentage.
```
import numpy as np
ozone_shots = (
shots
.assign(
is_goal=shots.event_type == "GOAL",
x=np.abs(shots.x),
y=shots.y * np.sign(shots.x),
)
)
fig, axs = plt.subplots(1, 3, figsize=(18, 8))
rink = NHLRink(rotation=270, net={"visible": False})
rink.contourf(
"x", "y", "is_goal", data=ozone_shots,
nbins=8, levels=30, plot_range="ozone", cmap="bwr",
ax=axs[0], draw_kw={"display_range": "ozone"},
)
rink.heatmap(
"x", "y", "is_goal", data=ozone_shots,
binsize=5, fill_value=0, plot_xlim=(25, 89), cmap="magma", vmax=0.25,
ax=axs[1], draw_kw={"display_range": "ozone"},
)
rink.hexbin(
"x", "y", "is_goal", data=ozone_shots,
gridsize=(14, 8), plot_range="ozone", alpha=0.85, vmax=0.25,
ax=axs[2], draw_kw={"display_range": "ozone"},
)
```
There's also a clear method which will attempt to remove anything that isn't part of the rink unless it's passed
to the keep variable. This can be useful for animations.
```
df = (
pd.read_csv("https://github.com/the-bucketless/bdc/raw/main/data/2022-02-08%20Canada%20at%20USA/2022-02-08%20Canada%20at%20USA%20P1%20PP1.csv")
.query("frame_id == 400")
.assign(team_color=lambda df_: np.where(df_.team_name == "Canada", "lightcoral", "aqua"))
)
rink = NHLRink(x_shift=100, y_shift=42.5, rotation=270)
fig, axs = plt.subplots(1, 2, figsize=(12, 8))
for ax in axs:
rink.draw(display_range="ozone", ax=ax)
rink.scatter(
"x_ft", "y_ft", ax=ax,
facecolor="team_color", edgecolor="black", s=300,
data=df,
)
rink.text(
"x_ft", "y_ft", "jersey_number", ax=ax,
ha="center", va="center", fontsize=14,
data=df,
)
teams_text = rink.text(
0.5, 0.05, "Canada vs USA", ax=ax,
use_rink_coordinates=False,
ha="center", va="center", fontsize=20,
)
rink.clear(ax=axs[1], keep=[teams_text])
```
## Inspiration
This project was partly inspired by [mplsoccer](https://github.com/andrewRowlinson/mplsoccer).
Hopefully, it can make things a little easier for anyone looking to get involved in hockey analytics.
## Contact
You can find me on twitter [@the_bucketless](https://twitter.com/the_bucketless) or email me at thebucketless@protonmail.com.
Raw data
{
"_id": null,
"home_page": "https://github.com/the-bucketless/hockey_rink",
"name": "hockey-rink",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "The Bucketless",
"author_email": "thebucketless@protonmail.com",
"download_url": "https://files.pythonhosted.org/packages/a5/25/0e2300bf143a8da1ab606e7a208b6e89b2521cd8e643bc675d0fc8bb9490/hockey_rink-1.0.1.tar.gz",
"platform": null,
"description": "\n\nA Python library for plotting hockey rinks with Matplotlib.\n\n## Installation\n```pip install hockey-rink```\n\n## Current Rinks\n\nThe following rinks are available for use:\n- Rink\n- NHLRink\n- NWHLRink (from the 2021 Lake Placid games)\n- IIHFRink\n- OldIIHFRink\n\n## Customization\nThere is also room for customization. For example, to change the dimension of a rink, update the length and/or width of the boards:\n\n```\nrink = Rink(boards={\"length\": 150, \"width\": 150, \"radius\": 75})\n```\n \n\n\nEach rink comes with a default set of features, but additional features can be added. Custom features should inherit \nfrom RinkFeature and override the _get_centered_xy method. The draw method can also be overridden if the desired feature can't be drawn\nwith a Matplotlib Polygon, though _get_centered_xy should still provide the feature's boundaries. The RinkImage\nfeature provides an example of this by inheriting from RinkRectangle.\n\nIf a custom feature is to be constrained to only display within the rink, the returned object needs to have a \nset_clip_path method.\n\n## Plots\nThere are currently wrappers available for the following Matplotlib plotting methods: \n- plot \n- scatter \n- arrow \n- hexbin \n- pcolormesh (heatmap in hockey-rink) \n- contour \n- contourf \n- text\n \nThere's also a rink.plot_fn which will take as its first argument a plotting method from either Matplotlib or seaborn \nand will attempt to make the desired plot.\n \n```\nfrom hockey_rink import NHLRink, RinkImage\nimport matplotlib.pyplot as plt\nimport pandas as pd\nimport seaborn as sns\n\nshots = (\n pd.read_parquet(\"https://github.com/sportsdataverse/fastRhockey-data/blob/main/nhl/pbp/parquet/play_by_play_2023.parquet?raw=true\")\n .query(\"event_type in ('GOAL', 'SHOT', 'MISS')\")\n)\n\nteam_colors = {\"San Jose Sharks\": (0, 0.5, 0.5), \"Nashville Predators\": (1, 0.7, 0.1)}\n\nrink = NHLRink(\n sharks_logo={\n \"feature_class\": RinkImage,\n \"image_path\": \"https://upload.wikimedia.org/wikipedia/en/thumb/3/37/SanJoseSharksLogo.svg/330px-SanJoseSharksLogo.svg.png\",\n \"x\": 55, \"length\": 50, \"width\": 42,\n \"zorder\": 15, \"alpha\": 0.5,\n },\n preds_logo={\n \"feature_class\": RinkImage,\n \"image_path\": \"https://upload.wikimedia.org/wikipedia/en/thumb/9/9c/Nashville_Predators_Logo_%282011%29.svg/330px-Nashville_Predators_Logo_%282011%29.svg.png\",\n \"x\": -55, \"length\": 50, \"width\": 29,\n \"zorder\": 15, \"alpha\": 0.5,\n }\n)\n\nfirst_period = shots.query(\"game_id == 2022020001 and period == 1\")\n\nfig, axs = plt.subplots(1, 2, figsize=(18, 8))\nrink.scatter(\"x\", \"y\", facecolor=first_period.event_team.map(team_colors), s=100, edgecolor=\"white\", data=first_period, ax=axs[0])\nrink.plot_fn(sns.scatterplot, x=\"x\", y=\"y\", hue=\"event_team\", s=100, legend=False, data=first_period, ax=axs[1], palette=team_colors);\n```\n\n\nWhen using plots that require binning, it's often best to include a plot_range even when it isn't being used to \nfind the bins. Here's an example using shooting percentage.\n\n```\nimport numpy as np\n\nozone_shots = (\n shots\n .assign(\n is_goal=shots.event_type == \"GOAL\",\n x=np.abs(shots.x),\n y=shots.y * np.sign(shots.x),\n )\n)\n\nfig, axs = plt.subplots(1, 3, figsize=(18, 8))\n\nrink = NHLRink(rotation=270, net={\"visible\": False})\n\nrink.contourf(\n \"x\", \"y\", \"is_goal\", data=ozone_shots, \n nbins=8, levels=30, plot_range=\"ozone\", cmap=\"bwr\",\n ax=axs[0], draw_kw={\"display_range\": \"ozone\"},\n)\n\nrink.heatmap(\n \"x\", \"y\", \"is_goal\", data=ozone_shots, \n binsize=5, fill_value=0, plot_xlim=(25, 89), cmap=\"magma\", vmax=0.25,\n ax=axs[1], draw_kw={\"display_range\": \"ozone\"},\n)\n\nrink.hexbin(\n \"x\", \"y\", \"is_goal\", data=ozone_shots,\n gridsize=(14, 8), plot_range=\"ozone\", alpha=0.85, vmax=0.25,\n ax=axs[2], draw_kw={\"display_range\": \"ozone\"},\n)\n```\n\n\nThere's also a clear method which will attempt to remove anything that isn't part of the rink unless it's passed \nto the keep variable. This can be useful for animations.\n```\ndf = (\n pd.read_csv(\"https://github.com/the-bucketless/bdc/raw/main/data/2022-02-08%20Canada%20at%20USA/2022-02-08%20Canada%20at%20USA%20P1%20PP1.csv\")\n .query(\"frame_id == 400\")\n .assign(team_color=lambda df_: np.where(df_.team_name == \"Canada\", \"lightcoral\", \"aqua\"))\n)\n\nrink = NHLRink(x_shift=100, y_shift=42.5, rotation=270)\n\nfig, axs = plt.subplots(1, 2, figsize=(12, 8))\nfor ax in axs:\n rink.draw(display_range=\"ozone\", ax=ax)\n \n rink.scatter(\n \"x_ft\", \"y_ft\", ax=ax,\n facecolor=\"team_color\", edgecolor=\"black\", s=300,\n data=df,\n )\n\n rink.text(\n \"x_ft\", \"y_ft\", \"jersey_number\", ax=ax,\n ha=\"center\", va=\"center\", fontsize=14, \n data=df,\n )\n \n teams_text = rink.text(\n 0.5, 0.05, \"Canada vs USA\", ax=ax,\n use_rink_coordinates=False,\n ha=\"center\", va=\"center\", fontsize=20,\n )\n\nrink.clear(ax=axs[1], keep=[teams_text])\n```\n\n\n## Inspiration\nThis project was partly inspired by [mplsoccer](https://github.com/andrewRowlinson/mplsoccer).\n\nHopefully, it can make things a little easier for anyone looking to get involved in hockey analytics.\n\n## Contact\nYou can find me on twitter [@the_bucketless](https://twitter.com/the_bucketless) or email me at thebucketless@protonmail.com.",
"bugtrack_url": null,
"license": "GNU General Public License v3 (GPLv3)",
"summary": "A Python library for plotting hockey rinks with Matplotlib.",
"version": "1.0.1",
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "a5250e2300bf143a8da1ab606e7a208b6e89b2521cd8e643bc675d0fc8bb9490",
"md5": "05801de94f1d413124c64b424bd8c23e",
"sha256": "a0003bf50a6df024bc9e3c0b7cd84ed0e2c79d5d53f4e5d809bb696d8c035d96"
},
"downloads": -1,
"filename": "hockey_rink-1.0.1.tar.gz",
"has_sig": false,
"md5_digest": "05801de94f1d413124c64b424bd8c23e",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 39838,
"upload_time": "2023-03-18T21:27:49",
"upload_time_iso_8601": "2023-03-18T21:27:49.930934Z",
"url": "https://files.pythonhosted.org/packages/a5/25/0e2300bf143a8da1ab606e7a208b6e89b2521cd8e643bc675d0fc8bb9490/hockey_rink-1.0.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-03-18 21:27:49",
"github": true,
"gitlab": false,
"bitbucket": false,
"github_user": "the-bucketless",
"github_project": "hockey_rink",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"requirements": [
{
"name": "matplotlib",
"specs": [
[
">=",
"3.3.4"
]
]
},
{
"name": "numpy",
"specs": [
[
">=",
"1.19.5"
]
]
},
{
"name": "pillow",
"specs": [
[
">=",
"8.2.0"
]
]
}
],
"lcname": "hockey-rink"
}
JSON
