Python时间序列数据分析与可视化实战指南

bash1pip install pandas matplotlib seaborn statsmodels yfinance

python1import yfinance as yf
2
3# 下载数据（参数：股票代码，时间范围）
4data = yf.download('600519.SS', start='2020-11-17', end='2025-11-17')
5# 提取收盘价作为分析对象
6close_prices = data['Close']
7print(close_prices.head())

12020-11-17    1750.00
22020-11-18    1735.50
32020-11-19    1748.80
4...

python1# 检查缺失值数量
2print(f"缺失值数量：{close_prices.isnull().sum()}")
3
4# 线性插值填充缺失值
5close_prices = close_prices.interpolate(method='linear')

python1# 确保索引为datetime类型
2close_prices.index = pd.to_datetime(close_prices.index)
3
4# 重采样为周数据（可选）
5weekly_data = close_prices.resample('W').last()

python1import matplotlib.pyplot as plt
2plt.figure(figsize=(14, 6))
3plt.plot(close_prices.index, close_prices.values, 
4         label='贵州茅台收盘价', color='#1f77b4')
5plt.title('2020-2025年贵州茅台股价走势', fontsize=16)
6plt.xlabel('日期', fontsize=12)
7plt.ylabel('价格（元）', fontsize=12)
8plt.grid(True, linestyle='--', alpha=0.7)
9plt.legend()
10plt.tight_layout()
11plt.show()

python1# 计算20日和60日移动平均
2ma_20 = close_prices.rolling(window=20).mean()
3ma_60 = close_prices.rolling(window=60).mean()
4
5plt.figure(figsize=(14, 6))
6plt.plot(close_prices.index, close_prices, label='收盘价', alpha=0.5)
7plt.plot(close_prices.index, ma_20, label='20日均线', linewidth=2)
8plt.plot(close_prices.index, ma_60, label='60日均线', linewidth=2)
9plt.title('股价与移动平均线对比', fontsize=16)
10plt.legend()
11plt.show()

python1from statsmodels.tsa.seasonal import seasonal_decompose
2
3# 按年周期分解（假设数据有年度季节性）
4result = seasonal_decompose(close_prices, model='additive', period=252)  # 252个交易日≈1年
5
6result.plot()
7plt.suptitle('股价季节性分解', y=1.02)
8plt.tight_layout()
9plt.show()

python1from statsmodels.graphics.tsaplots import plot_acf
2
3plt.figure(figsize=(12, 6))
4plot_acf(close_prices.dropna(), lags=60, alpha=0.05)  # 显示60阶自相关
5plt.title('股价自相关图', fontsize=16)
6plt.xlabel('滞后阶数（交易日）', fontsize=12)
7plt.ylabel('自相关系数', fontsize=12)
8plt.show()

bash1pip install plotly

python1import plotly.graph_objects as go
2
3fig = go.Figure()
4fig.add_trace(go.Scatter(
5    x=close_prices.index,
6    y=close_prices,
7    name='收盘价',
8    line=dict(color='#1f77b4', width=2)
9))
10
11# 添加移动平均线
12fig.add_trace(go.Scatter(
13    x=ma_20.index,
14    y=ma_20,
15    name='20日均线',
16    line=dict(color='#ff7f0e', width=1.5, dash='dash')
17))
18
19fig.update_layout(
20    title='贵州茅台股价交互式图表',
21    xaxis_title='日期',
22    yaxis_title='价格（元）',
23    hovermode='x unified',
24    template='plotly_white'
25)
26
27fig.show()

python1# 计算滚动均值和标准差
2rolling_mean = close_prices.rolling(window=20).mean()
3rolling_std = close_prices.rolling(window=20).std()
4
5# 定义异常阈值
6upper_bound = rolling_mean + 3 * rolling_std
7lower_bound = rolling_mean - 3 * rolling_std
8
9# 检测异常值
10anomalies = close_prices[(close_prices > upper_bound) | (close_prices < lower_bound)]
11
12# 可视化
13plt.figure(figsize=(14, 6))
14plt.plot(close_prices.index, close_prices, label='收盘价', alpha=0.7)
15plt.plot(upper_bound.index, upper_bound, 'r--', label='上界')
16plt.plot(lower_bound.index, lower_bound, 'r--', label='下界')
17plt.scatter(anomalies.index, anomalies, color='red', label='异常值', zorder=5)
18plt.title('股价异常值检测（3σ原则）', fontsize=16)
19plt.legend()
20plt.show()

python1# 生成完整日期范围
2full_dates = pd.date_range(start='2020-11-17', end='2025-11-17', freq='B')  # B表示工作日
3close_prices = close_prices.reindex(full_dates)
4close_prices = close_prices.interpolate(method='linear')  # 线性插值填充

python1plt.savefig('stock_analysis.png', dpi=300, bbox_inches='tight')

python1fig.write_html('interactive_chart.html')  # 保存为HTML
2fig.write_image("interactive_chart.png", scale=2)  # 需要安装kaleido库

python1# 获取多只股票数据
2stocks = yf.download(['600519.SS', '000858.SZ', '601318.SH'], 
3                    start='2020-11-17', end='2025-11-17')['Close']
4
5# 绘制子图
6fig, axes = plt.subplots(3, 1, figsize=(14, 12))
7for i, code in enumerate(stocks.columns):
8    axes[i].plot(stocks.index, stocks[code], label=code)
9    axes[i].set_title(f'{code}股价走势')
10    axes[i].legend()
11plt.tight_layout()
12plt.show()