matplotlib进阶教程：如何逐步美化一个折线图

大家好，今天分享一个非常有趣的 Python 教程，如何美化一个 matplotlib 折线图，喜欢记得收藏、关注、点赞。

注：数据、完整代码、技术交流文末获取

1. 导入包

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
import matplotlib.gridspec as gridspec

2. 获得数据

file_id = '1yM_F93NY4QkxjlKL3GzdcCQEnBiA2ltB'
url = f'https://drive.google.com/uc?id={file_id}'
df = pd.read_csv(url, index_col=0)
df

数据长得是这样的：

3. 对数据做一些预处理

按照需要，对数据再做一些预处理，代码及效果如下：

home_df = df.copy()
home_df = home_df.melt(id_vars = ["date", "home_team_name", "away_team_name"])
home_df["venue"] = "H"
home_df.rename(columns = {"home_team_name":"team", "away_team_name":"opponent"}, inplace = True)
home_df.replace({"variable":{"home_team_xG":"xG_for", "away_team_xG":"xG_ag"}}, inplace = True)

away_df = df.copy()
away_df = away_df.melt(id_vars = ["date", "away_team_name", "home_team_name"])
away_df["venue"] = "A"
away_df.rename(columns = {"away_team_name":"team", "home_team_name":"opponent"}, inplace = True)
away_df.replace({"variable":{"away_team_xG":"xG_for", "home_team_xG":"xG_ag"}}, inplace = True)

df = pd.concat([home_df, away_df]).reset_index(drop = True)
df

4. 画图

# ---- Filter the data

Y_for = df[(df["team"] == "Lazio") & (df["variable"] == "xG_for")]["value"].reset_index(drop = True)
Y_ag = df[(df["team"] == "Lazio") & (df["variable"] == "xG_ag")]["value"].reset_index(drop = True)
X_ = pd.Series(range(len(Y_for)))

# ---- Compute rolling average

Y_for = Y_for.rolling(window = 5, min_periods = 0).mean() # min_periods is for partial avg.
Y_ag = Y_ag.rolling(window = 5, min_periods = 0).mean()

fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

ax.plot(X_, Y_for)
ax.plot(X_, Y_ag)

使用matplotlib倒是可以快速把图画好了，但是太丑了。接下来进行优化。

4.1 优化：添加点

这里为每一个数据添加点

fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

# --- Remove spines and add gridlines

ax.spines["left"].set_visible(False)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)

ax.grid(ls = "--", lw = 0.5, color = "#4E616C")

# --- The data

ax.plot(X_, Y_for, marker = "o")
ax.plot(X_, Y_ag, marker = "o")

4.2 优化：设置刻度

fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

# --- Remove spines and add gridlines

ax.spines["left"].set_visible(False)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)

ax.grid(ls = "--", lw = 0.25, color = "#4E616C")

# --- The data

ax.plot(X_, Y_for, marker = "o", mfc = "white", ms = 5)
ax.plot(X_, Y_ag, marker = "o", mfc = "white", ms = 5)

# --- Adjust tickers and spine to match the style of our grid

ax.xaxis.set_major_locator(ticker.MultipleLocator(2)) # ticker every 2 matchdays
xticks_ = ax.xaxis.set_ticklabels([x - 1 for x in range(0, len(X_) + 3, 2)])
# This last line outputs
# [-1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33, 35]
# and we mark the tickers every two positions.

ax.xaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)
ax.yaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)

ax.spines["bottom"].set_edgecolor("#4E616C")

4.3 优化：设置填充

fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

# --- Remove spines and add gridlines

ax.spines["left"].set_visible(False)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)

ax.grid(ls = "--", lw = 0.25, color = "#4E616C")

# --- The data

ax.plot(X_, Y_for, marker = "o", mfc = "white", ms = 5)
ax.plot(X_, Y_ag, marker = "o", mfc = "white", ms = 5)

# --- Fill between

ax.fill_between(x = X_, y1 = Y_for, y2 = Y_ag, alpha = 0.5)

# --- Adjust tickers and spine to match the style of our grid

ax.xaxis.set_major_locator(ticker.MultipleLocator(2)) # ticker every 2 matchdays
xticks_ = ax.xaxis.set_ticklabels([x - 1 for x in range(0, len(X_) + 3, 2)])

ax.xaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)
ax.yaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)

ax.spines["bottom"].set_edgecolor("#4E616C")

4.4 优化：设置填充颜色

1. 当橙色线更高时，希望填充为橙色。但是上面的还无法满足，这里再优化一下.

fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

# --- Remove spines and add gridlines

ax.spines["left"].set_visible(False)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)

ax.grid(ls = "--", lw = 0.25, color = "#4E616C")

# --- The data

ax.plot(X_, Y_for, marker = "o", mfc = "white", ms = 5)
ax.plot(X_, Y_ag, marker = "o", mfc = "white", ms = 5)

# --- Fill between

# Identify points where Y_for > Y_ag

pos_for = (Y_for > Y_ag)
ax.fill_between(x = X_[pos_for], y1 = Y_for[pos_for], y2 = Y_ag[pos_for], alpha = 0.5)

pos_ag = (Y_for <= Y_ag)
ax.fill_between(x = X_[pos_ag], y1 = Y_for[pos_ag], y2 = Y_ag[pos_ag], alpha = 0.5)

# --- Adjust tickers and spine to match the style of our grid

ax.xaxis.set_major_locator(ticker.MultipleLocator(2)) # ticker every 2 matchdays
xticks_ = ax.xaxis.set_ticklabels([x - 1 for x in range(0, len(X_) + 3, 2)])

ax.xaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)
ax.yaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)

ax.spines["bottom"].set_edgecolor("#4E616C")

上面的图出现异常，再修改一下：

X_aux = X_.copy()
X_aux.index = X_aux.index * 10 # 9 aux points in between each match
last_idx = X_aux.index[-1] + 1
X_aux = X_aux.reindex(range(last_idx))
X_aux = X_aux.interpolate()


# --- Aux series for the xG created (Y_for)
Y_for_aux = Y_for.copy()
Y_for_aux.index = Y_for_aux.index * 10
last_idx = Y_for_aux.index[-1] + 1
Y_for_aux = Y_for_aux.reindex(range(last_idx))
Y_for_aux = Y_for_aux.interpolate()

# --- Aux series for the xG conceded (Y_ag)
Y_ag_aux = Y_ag.copy()
Y_ag_aux.index = Y_ag_aux.index * 10
last_idx = Y_ag_aux.index[-1] + 1
Y_ag_aux = Y_ag_aux.reindex(range(last_idx))
Y_ag_aux = Y_ag_aux.interpolate()



fig, ax = plt.subplots(figsize = (7,3), dpi = 200)

# --- Remove spines and add gridlines

ax.spines["left"].set_visible(False)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)

ax.grid(ls = "--", lw = 0.25, color = "#4E616C")

# --- The data

for_ = ax.plot(X_, Y_for, marker = "o", mfc = "white", ms = 5)
ag_ = ax.plot(X_, Y_ag, marker = "o", mfc = "white", ms = 5)

# --- Fill between

for index in range(len(X_aux) - 1):
    # Choose color based on which line's on top
    if Y_for_aux.iloc[index + 1] > Y_ag_aux.iloc[index + 1]:
        color = for_[0].get_color()
    else:
        color = ag_[0].get_color()
    
    # Fill between the current point and the next point in pur extended series.
    ax.fill_between([X_aux[index], X_aux[index+1]], 
                    [Y_for_aux.iloc[index], Y_for_aux.iloc[index+1]], 
                    [Y_ag_aux.iloc[index], Y_ag_aux.iloc[index+1]], 
                    color=color, zorder = 2, alpha = 0.2, ec = None)

# --- Adjust tickers and spine to match the style of our grid

ax.xaxis.set_major_locator(ticker.MultipleLocator(2)) # ticker every 2 matchdays
xticks_ = ax.xaxis.set_ticklabels([x - 1 for x in range(0, len(X_) + 3, 2)])

ax.xaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)
ax.yaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)

ax.spines["bottom"].set_edgecolor("#4E616C")

5. 把功能打包成函数

1. 上面的样子都还不错啦，接下来把这些东西都打包成一个函数。方便后面直接出图。

def plot_xG_rolling(team, ax, window = 5, color_for = "blue", color_ag = "orange", data = df):
  '''
  This function creates a rolling average xG plot for a given team and rolling
  window.

  team (str): The team's name
  ax (obj): a Matplotlib axes.
  window (int): The number of periods for our rolling average.
  color_for (str): A hex color code for xG created.
  color_af (str): A hex color code for xG conceded.
  data (DataFrame): our df with the xG data.
  '''

  # -- Prepping the data
  home_df = data.copy()
  home_df = home_df.melt(id_vars = ["date", "home_team_name", "away_team_name"])
  home_df["venue"] = "H"
  home_df.rename(columns = {"home_team_name":"team", "away_team_name":"opponent"}, inplace = True)
  home_df.replace({"variable":{"home_team_xG":"xG_for", "away_team_xG":"xG_ag"}}, inplace = True)

  away_df = data.copy()
  away_df = away_df.melt(id_vars = ["date", "away_team_name", "home_team_name"])
  away_df["venue"] = "A"
  away_df.rename(columns = {"away_team_name":"team", "home_team_name":"opponent"}, inplace = True)
  away_df.replace({"variable":{"away_team_xG":"xG_for", "home_team_xG":"xG_ag"}}, inplace = True)

  df = pd.concat([home_df, away_df]).reset_index(drop = True)

  # ---- Filter the data

  Y_for = df[(df["team"] == team) & (df["variable"] == "xG_for")]["value"].reset_index(drop = True)
  Y_ag = df[(df["team"] == team) & (df["variable"] == "xG_ag")]["value"].reset_index(drop = True)
  X_ = pd.Series(range(len(Y_for)))

  if Y_for.shape[0] == 0:
    raise ValueError(f"Team {team} is not present in the DataFrame")

  # ---- Compute rolling average

  Y_for = Y_for.rolling(window = 5, min_periods = 0).mean() # min_periods is for partial avg.
  Y_ag = Y_ag.rolling(window = 5, min_periods = 0).mean()

  # ---- Create auxiliary series for filling between curves

  X_aux = X_.copy()
  X_aux.index = X_aux.index * 10 # 9 aux points in between each match
  last_idx = X_aux.index[-1] + 1
  X_aux = X_aux.reindex(range(last_idx))
  X_aux = X_aux.interpolate()

  # --- Aux series for the xG created (Y_for)
  Y_for_aux = Y_for.copy()
  Y_for_aux.index = Y_for_aux.index * 10
  last_idx = Y_for_aux.index[-1] + 1
  Y_for_aux = Y_for_aux.reindex(range(last_idx))
  Y_for_aux = Y_for_aux.interpolate()

  # --- Aux series for the xG conceded (Y_ag)
  Y_ag_aux = Y_ag.copy()
  Y_ag_aux.index = Y_ag_aux.index * 10
  last_idx = Y_ag_aux.index[-1] + 1
  Y_ag_aux = Y_ag_aux.reindex(range(last_idx))
  Y_ag_aux = Y_ag_aux.interpolate()

  # --- Plotting our data

  # --- Remove spines and add gridlines

  ax.spines["left"].set_visible(False)
  ax.spines["top"].set_visible(False)
  ax.spines["right"].set_visible(False)

  ax.grid(ls = "--", lw = 0.25, color = "#4E616C")

  # --- The data

  for_ = ax.plot(X_, Y_for, marker = "o", mfc = "white", ms = 4, color = color_for)
  ag_ = ax.plot(X_, Y_ag, marker = "o", mfc = "white", ms = 4, color = color_ag)

  # --- Fill between

  for index in range(len(X_aux) - 1):
      # Choose color based on which line's on top
      if Y_for_aux.iloc[index + 1] > Y_ag_aux.iloc[index + 1]:
          color = for_[0].get_color()
      else:
          color = ag_[0].get_color()
      
      # Fill between the current point and the next point in pur extended series.
      ax.fill_between([X_aux[index], X_aux[index+1]], 
                      [Y_for_aux.iloc[index], Y_for_aux.iloc[index+1]], 
                      [Y_ag_aux.iloc[index], Y_ag_aux.iloc[index+1]], 
                      color=color, zorder = 2, alpha = 0.2, ec = None)
      

  # --- Ensure minimum value of Y-axis is zero
  ax.set_ylim(0)

  # --- Adjust tickers and spine to match the style of our grid

  ax.xaxis.set_major_locator(ticker.MultipleLocator(2)) # ticker every 2 matchdays
  xticks_ = ax.xaxis.set_ticklabels([x - 1 for x in range(0, len(X_) + 3, 2)])

  ax.xaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)
  ax.yaxis.set_tick_params(length = 2, color = "#4E616C", labelcolor = "#4E616C", labelsize = 6)

  ax.spines["bottom"].set_edgecolor("#4E616C")

  # --- Legend and team name

  Y_for_last = Y_for.iloc[-1]
  Y_ag_last = Y_ag.iloc[-1]

  # -- Add the team's name
  team_ = ax.text(
            x = 0, y = ax.get_ylim()[1] + ax.get_ylim()[1]/20,
            s = f'{team}',
            color = "#4E616C",
            va = 'center',
            ha = 'left',
            size = 7
          )
  
  # -- Add the xG created label
  for_label_ = ax.text(
            x = X_.iloc[-1] + 0.75, y = Y_for_last,
            s = f'{Y_for_last:,.1f} xGF',
            color = color_for,
            va = 'center',
            ha = 'left',
            size = 6.5
          )

  # -- Add the xG conceded label
  ag_label_ = ax.text(
            x = X_.iloc[-1] + 0.75, y = Y_ag_last,
            s = f'{Y_ag_last:,.1f} xGA',
            color = color_ag,
            va = 'center',
            ha = 'left',
            size = 6.5
          )

6.1 测试函数

file_id = '1yM_F93NY4QkxjlKL3GzdcCQEnBiA2ltB'
url = f'https://drive.google.com/uc?id={file_id}'
df = pd.read_csv(url, index_col=0)

fig = plt.figure(figsize=(5, 2), dpi = 200)
ax = plt.subplot(111)

plot_xG_rolling("Sassuolo", ax, color_for = "#00A752", color_ag = "black", data = df)

plt.tight_layout()

再设置更加丰富的颜色：

fig = plt.figure(figsize=(5, 8), dpi = 200, facecolor = "#EFE9E6")

ax1 = plt.subplot(411, facecolor = "#EFE9E6")
ax2 = plt.subplot(412, facecolor = "#EFE9E6")
ax3 = plt.subplot(413, facecolor = "#EFE9E6")
ax4 = plt.subplot(414, facecolor = "#EFE9E6")

plot_xG_rolling("Sassuolo", ax1, color_for = "#00A752", color_ag = "black", data = df)
plot_xG_rolling("Lazio", ax2, color_for = "#87D8F7", color_ag = "#15366F", data = df)
plot_xG_rolling("Hellas Verona", ax3, color_for = "#153aab", color_ag = "#fdcf41", data = df)
plot_xG_rolling("Empoli", ax4, color_for = "#00579C", color_ag = "black", data = df)

plt.tight_layout()

最后

其实本文主要是对两个折线图做了一系列的优化和改进而已，主要是强调细节部分。

涉及到的matplotlib的知识，也主要是在ticks、背景颜色、fill_between部分。

推荐文章

• 李宏毅《机器学习》国语课程(2022)来了

• 有人把吴恩达老师的机器学习和深度学习做成了中文版

• 上瘾了，最近又给公司撸了一个可视化大屏(附源码)

• 如此优雅，4款 Python 自动数据分析神器真香啊

• 梳理半月有余，精心准备了17张知识思维导图，这次要讲清统计学

• 年终汇总：20份可视化大屏模板，直接套用真香(文末附源码)

技术交流

欢迎转载、收藏、有所收获点赞支持一下！数据、代码可以找我获取

目前开通了技术交流群，群友已超过2000人，添加时最好的备注方式为：来源+兴趣方向，方便找到志同道合的朋友

• 方式①、发送如下图片至微信，长按识别，后台回复：加群；
• 方式②、添加微信号：dkl88191，备注：来自CSDN
• 方式③、微信搜索公众号：Python学习与数据挖掘，后台回复：加群