サンプル問題３－２　データ分析（サッカーデータの回帰分析）

# 必要なライブラリをインポート
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression

# データをファイルパスから直接読み込む
file_path = 'J1league_all_2023_ENG.csv'  # ここにファイルパスを指定してください
data = pd.read_csv(file_path)

# 列名を英語にリネーム
data = data.rename(columns={
    'チーム名': 'Team Name',
    '1試合平均得点数': 'Average Goals per Game',
    '1試合平均パス数': 'Average Passes per Game',
    'トップ9': 'Top 9'
})

# 必要なカラムのみフィルタリング
data_filtered = data[['Team Name', 'Average Goals per Game', 'Average Passes per Game', 'Top 9']]

# トップ9チームとその他のチームのデータを分ける
top_9_teams = data_filtered[data_filtered["Top 9"] == 1]
other_teams = data_filtered[data_filtered["Top 9"] == 0]

# 回帰直線を計算する関数
def calculate_regression(X, y):
    X_reshaped = X.values.reshape(-1, 1)  # Xを2次元配列に変換
    model = LinearRegression().fit(X_reshaped, y)  # 回帰モデルをフィッティング
    regression_line = model.predict(np.linspace(X.min(), X.max(), 100).reshape(-1, 1))  # 回帰直線を計算
    return model, regression_line

# トップ9チームの回帰直線と相関係数を計算
model_top9, line_top9 = calculate_regression(top_9_teams["Average Passes per Game"], top_9_teams["Average Goals per Game"])
correlation_top9 = np.corrcoef(top_9_teams["Average Passes per Game"], top_9_teams["Average Goals per Game"])[0, 1]

# その他のチームの回帰直線と相関係数を計算
model_other, line_other = calculate_regression(other_teams["Average Passes per Game"], other_teams["Average Goals per Game"])
correlation_other = np.corrcoef(other_teams["Average Passes per Game"], other_teams["Average Goals per Game"])[0, 1]

# トップ9チーム用のグラフを描画
plt.figure(figsize=(10, 6))
plt.scatter(top_9_teams["Average Passes per Game"], top_9_teams["Average Goals per Game"], color='blue', label='Top 9 Teams')

# 回帰直線を描画
plt.plot(np.linspace(top_9_teams["Average Passes per Game"].min(), top_9_teams["Average Passes per Game"].max(), 100), line_top9, color='blue',
         label=f'Regression: y = {model_top9.coef_[0]:.4f}x + {model_top9.intercept_:.2f}, Corr = {correlation_top9:.4f}')

# チーム名をマーカーの上部に表示
for i, team in enumerate(top_9_teams['Team Name']):
    plt.text(top_9_teams["Average Passes per Game"].values[i], 
             top_9_teams["Average Goals per Game"].values[i] + 0.02,  # 少し上に配置するために +0.02 を追加
             team, fontsize=9, color='blue')

# グラフの設定
plt.title('Average Goals per Game vs Average Passes per Game (Top 9 Teams)')
plt.xlabel('Average Passes per Game')
plt.ylabel('Average Goals per Game')
plt.legend()
plt.grid(True)
plt.show()

# その他のチーム用のグラフを描画
plt.figure(figsize=(10, 6))
plt.scatter(other_teams["Average Passes per Game"], other_teams["Average Goals per Game"], color='red', label='Other Teams')

# 回帰直線を描画
plt.plot(np.linspace(other_teams["Average Passes per Game"].min(), other_teams["Average Passes per Game"].max(), 100), line_other, color='red',
         label=f'Regression: y = {model_other.coef_[0]:.4f}x + {model_other.intercept_:.2f}, Corr = {correlation_other:.4f}')

# チーム名をマーカーの上部に表示
for i, team in enumerate(other_teams['Team Name']):
    plt.text(other_teams["Average Passes per Game"].values[i], 
             other_teams["Average Goals per Game"].values[i] + 0.02,  # 少し上に配置するために +0.02 を追加
             team, fontsize=9, color='red')

# グラフの設定
plt.title('Average Goals per Game vs Average Passes per Game (Other Teams)')
plt.xlabel('Average Passes per Game')
plt.ylabel('Average Goals per Game')
plt.legend()
plt.grid(True)
plt.show()