如何使用 Python 建立一個 NBA 得分圖？

時間 2019-12-09

原文原文鏈接

本文意在建立一個得分圖，該圖同時描繪了從場上不一樣位置投籃得分的百分比和投籃次數，這和 Austin Clemen 我的網站上的帖子 http://www.austinclemens.com/shotcharts/ 相似。html

爲了實現這個得分圖，筆者參考了 Savvas Tjortjoglou 的帖子 http://savvastjortjoglou.com/nba-shot-sharts.html。這篇帖子很棒，可是他只描述了從不一樣位置投籃的次數。而筆者對在不一樣位置的投籃次數和進球百分比都很感興趣，因此還須要進一步的工做，在原有基礎上添加些東西，下面是實現過程。python

#import some libraries and tell ipython we want inline figures rather than interactive figures. 
%matplotlib inline
import matplotlib.pyplot as plt, pandas as pd, numpy as np, matplotlib as mpl

首先，咱們須要得到每一個球員的投籃數據。利用 Savvas Tjortjoglou 貼出的代碼，筆者從 NBA.com 網站 API 上獲取了數據。在此不會貼出這個函數的結果。若是你感興趣，推薦你去看看 Savvas Tjortjoglou 的博客。數據庫

def aqcuire_shootingData(PlayerID,Season):
    import requests
    shot_chart_url = 'http://stats.nba.com/stats/shotchartdetail?CFID=33&CFPARAMS='+Season+'&ContextFilter='\
                    '&ContextMeasure=FGA&DateFrom=&DateTo=&GameID=&GameSegment=&LastNGames=0&LeagueID='\
                    '00&Location=&MeasureType=Base&Month=0&OpponentTeamID=0&Outcome=&PaceAdjust='\
                    'N&PerMode=PerGame&Period=0&PlayerID='+PlayerID+'&PlusMinus=N&Position=&Rank='\
                    'N&RookieYear=&Season='+Season+'&SeasonSegment=&SeasonType=Regular+Season&TeamID='\
                    '0&VsConference=&VsDivision=&mode=Advanced&showDetails=0&showShots=1&showZones=0'
    response = requests.get(shot_chart_url)
    headers = response.json()['resultSets'][0]['headers']
    shots = response.json()['resultSets'][0]['rowSet']
    shot_df = pd.DataFrame(shots, columns=headers)
    return shot_df

接下來，咱們須要繪製一個包含得分圖的籃球場圖。該籃球場圖例必須使用與NBA.com API 相同的座標系統。例如，3分位置的投籃距籃筐必須爲 X 單位，上籃距離籃筐則是 Y 單位。一樣，筆者再次使用了 Savvas Tjortjoglou 的代碼（哈哈，不然的話，搞明白 NBA.com 網站的座標系統確定會耗費很多的時間）。json

def draw_court(ax=None, color='black', lw=2, outer_lines=False):
    from matplotlib.patches import Circle, Rectangle, Arc
    if ax is None:
        ax = plt.gca()
    hoop = Circle((0, 0), radius=7.5, linewidth=lw, color=color, fill=False)
    backboard = Rectangle((-30, -7.5), 60, -1, linewidth=lw, color=color)
    outer_box = Rectangle((-80, -47.5), 160, 190, linewidth=lw, color=color,
                          fill=False)
    inner_box = Rectangle((-60, -47.5), 120, 190, linewidth=lw, color=color,
                          fill=False)
    top_free_throw = Arc((0, 142.5), 120, 120, theta1=0, theta2=180,
                         linewidth=lw, color=color, fill=False)
    bottom_free_throw = Arc((0, 142.5), 120, 120, theta1=180, theta2=0,
                            linewidth=lw, color=color, linestyle='dashed')
    restricted = Arc((0, 0), 80, 80, theta1=0, theta2=180, linewidth=lw,
                     color=color)
    corner_three_a = Rectangle((-220, -47.5), 0, 140, linewidth=lw,
                               color=color)
    corner_three_b = Rectangle((220, -47.5), 0, 140, linewidth=lw, color=color)
    three_arc = Arc((0, 0), 475, 475, theta1=22, theta2=158, linewidth=lw,
                    color=color)
    center_outer_arc = Arc((0, 422.5), 120, 120, theta1=180, theta2=0,
                           linewidth=lw, color=color)
    center_inner_arc = Arc((0, 422.5), 40, 40, theta1=180, theta2=0,
                           linewidth=lw, color=color)
    court_elements = [hoop, backboard, outer_box, inner_box, top_free_throw,
                      bottom_free_throw, restricted, corner_three_a,
                      corner_three_b, three_arc, center_outer_arc,
                      center_inner_arc]
    if outer_lines:
        outer_lines = Rectangle((-250, -47.5), 500, 470, linewidth=lw,
                                color=color, fill=False)
        court_elements.append(outer_lines)

    for element in court_elements:
        ax.add_patch(element)

    ax.set_xticklabels([])
    ax.set_yticklabels([])
    ax.set_xticks([])
    ax.set_yticks([])
    return ax

我想創造一個不一樣位置的投籃百分比數組，所以決定利用 matplot 的 Hexbin 函數 http://matplotlib.org/api/pyplot_api.html 將投籃位置均勻地分組到六邊形中。該函數會對每一個六邊形中每個位置的投籃次數進行計數。api

六邊形是均勻的分佈在 XY 網格中。「gridsize」變量控制六邊形的數目。「extent」變量控制第一個和最後一個六邊形的繪製位置（通常來講第一個六邊形的位置基於第一個投籃的位置）。數組

計算命中率則須要對每一個六邊形中投籃的次數和投籃得分次數進行計數，所以筆者對同一位置的投籃和得分數分別運行 hexbin 函數。而後，只需用每一個位置的進球數除以投籃數。服務器

def find_shootingPcts(shot_df, gridNum):
    x = shot_df.LOC_X[shot_df['LOC_Y']<425.1] #i want to make sure to only include shots I can draw
    y = shot_df.LOC_Y[shot_df['LOC_Y']<425.1]

    x_made = shot_df.LOC_X[(shot_df['SHOT_MADE_FLAG']==1) & (shot_df['LOC_Y']<425.1)]
    y_made = shot_df.LOC_Y[(shot_df['SHOT_MADE_FLAG']==1) & (shot_df['LOC_Y']<425.1)]

    #compute number of shots made and taken from each hexbin location
    hb_shot = plt.hexbin(x, y, gridsize=gridNum, extent=(-250,250,425,-50));
    plt.close() #don't want to show this figure!
    hb_made = plt.hexbin(x_made, y_made, gridsize=gridNum, extent=(-250,250,425,-50),cmap=plt.cm.Reds);
    plt.close()

    #compute shooting percentage
    ShootingPctLocs = hb_made.get_array() / hb_shot.get_array()
    ShootingPctLocs[np.isnan(ShootingPctLocs)] = 0 #makes 0/0s=0
    return (ShootingPctLocs, hb_shot)

筆者很是喜歡 Savvas Tjortjoglou 在他的得分圖中加入了球員頭像的作法，所以也順道用了他的這部分代碼。球員照片會出如今得分圖的右下角。app

def acquire_playerPic(PlayerID, zoom, offset=(250,400)):
    from matplotlib import  offsetbox as osb
    import urllib
    pic = urllib.urlretrieve("http://stats.nba.com/media/players/230x185/"+PlayerID+".png",PlayerID+".png")
    player_pic = plt.imread(pic[0])
    img = osb.OffsetImage(player_pic, zoom)
    #img.set_offset(offset)
    img = osb.AnnotationBbox(img, offset,xycoords='data',pad=0.0, box_alignment=(1,0), frameon=False)
    return img

筆者想用連續的顏色圖來描述投籃進球百分比，紅圈越多表明着更高的進球百分比。雖然「紅」顏色圖示效果不錯，可是它會將0%的投籃進球百分比顯示爲白色 http://matplotlib.org/users/colormaps.html，而這樣顯示就會不明顯，因此筆者用淡粉紅色表明0%的命中率，所以對紅顏色圖作了下面的修改。函數

#cmap = plt.cm.Reds
#cdict = cmap._segmentdata
cdict = {
    'blue': [(0.0, 0.6313725709915161, 0.6313725709915161), (0.25, 0.4470588266849518, 0.4470588266849518), (0.5, 0.29019609093666077, 0.29019609093666077), (0.75, 0.11372549086809158, 0.11372549086809158), (1.0, 0.05098039284348488, 0.05098039284348488)],
    'green': [(0.0, 0.7333333492279053, 0.7333333492279053), (0.25, 0.572549045085907, 0.572549045085907), (0.5, 0.4156862795352936, 0.4156862795352936), (0.75, 0.0941176488995552, 0.0941176488995552), (1.0, 0.0, 0.0)],
    'red': [(0.0, 0.9882352948188782, 0.9882352948188782), (0.25, 0.9882352948188782, 0.9882352948188782), (0.5, 0.9843137264251709, 0.9843137264251709), (0.75, 0.7960784435272217, 0.7960784435272217), (1.0, 0.40392157435417175, 0.40392157435417175)]
}

mymap = mpl.colors.LinearSegmentedColormap('my_colormap', cdict, 1024)

好了，如今須要作的就是將它們合併到一起。下面所示的較大函數會利用上文描述的函數來建立一個描述投籃命中率的得分圖，百分比由紅圈表示（紅色越深 = 更高的命中率），投籃次數則由圓圈的大小決定（圓圈越大 = 投籃次數越多）。須要注意的是，圓圈在交疊以前都能增大。一旦圓圈開始交疊，就沒法繼續增大。oop

在這個函數中，計算了每一個位置的投籃進球百分比和投籃次數。而後畫出在該位置投籃的次數（圓圈大小）和進球百分比（圓圈顏色深淺）。

def shooting_plot(shot_df, plot_size=(12,8),gridNum=30):
    from matplotlib.patches import Circle
    x = shot_df.LOC_X[shot_df['LOC_Y']<425.1]
    y = shot_df.LOC_Y[shot_df['LOC_Y']<425.1]

    #compute shooting percentage and # of shots
    (ShootingPctLocs, shotNumber) = find_shootingPcts(shot_df, gridNum)

    #draw figure and court
    fig = plt.figure(figsize=plot_size)#(12,7)
    cmap = mymap #my modified colormap
    ax = plt.axes([0.1, 0.1, 0.8, 0.8]) #where to place the plot within the figure
    draw_court(outer_lines=False)
    plt.xlim(-250,250)
    plt.ylim(400, -25)

    #draw player image
    zoom = np.float(plot_size[0])/(12.0*2) #how much to zoom the player's pic. I have this hackily dependent on figure size
    img = acquire_playerPic(PlayerID, zoom)
    ax.add_artist(img)

    #draw circles
    for i, shots in enumerate(ShootingPctLocs):
        restricted = Circle(shotNumber.get_offsets()[i], radius=shotNumber.get_array()[i],
                            color=cmap(shots),alpha=0.8, fill=True)
        if restricted.radius > 240/gridNum: restricted.radius=240/gridNum
        ax.add_patch(restricted)

    #draw color bar
    ax2 = fig.add_axes([0.92, 0.1, 0.02, 0.8])
    cb = mpl.colorbar.ColorbarBase(ax2,cmap=cmap, orientation='vertical')
    cb.set_label('Shooting %')
    cb.set_ticks([0.0, 0.25, 0.5, 0.75, 1.0])
    cb.set_ticklabels(['0%','25%', '50%','75%', '100%'])

    plt.show()
    return ax

好了，大功告成！由於筆者是森林狼隊的粉絲，在下面用幾分鐘跑出了森林狼隊前六甲的得分圖。

PlayerID = '203952' #andrew wiggins
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

PlayerID = '1626157' #karl anthony towns
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

PlayerID = '203897' #zach lavine
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

PlayerID = '203476' #gorgui deing
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

PlayerID = '2755' #kevin martin
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

PlayerID = '201937' #ricky rubio
shot_df = aqcuire_shootingData(PlayerID,'2015-16')
ax = shooting_plot(shot_df, plot_size=(12,8));

使用 hexbin 函數也是有隱患的，第一它並無解釋因爲三分線而致使的非線性特性（一些 hexbin 函數同時包括了2分和3分的投籃）。它很好的限定了一些窗口來進行3分投籃，但若是沒有這個位置的硬編碼就沒有辦法作到這一點。此外 hexbin 方法的一個優勢與是能夠很容易地改變窗口的數量，但不肯定是否能夠一樣靈活的處理2分投籃和3分投籃。

另一個隱患在於此圖將全部投籃都一視同仁，這至關不公平。在禁區投籃命中40%和三分線後的投籃命中40%但是大不相同。Austin Clemens 的解決辦法是將命中率與聯賽平均分關聯。也許過幾天筆者也會實現與之相似的功能。

原文 Creating NBA Shot Charts 做者 Dan Vatterott ，本文由 OneAPM 工程師編譯整理。

OneAPM 可以幫你查看 Python 應用程序的方方面面，不只可以監控終端的用戶體驗，還能監控服務器性能，同時還支持追蹤數據庫、第三方 API 和 Web 服務器的各類問題。想閱讀更多技術文章，請訪問 OneAPM 官方技術博客。
本文轉自 OneAPM 官方博客