18.2 对比Search Console vs 生成引擎引用数据

# 示例：计算覆盖度差异
sc_queries = set(['如何优化LCP', 'SEO技术指南', 'Core Web Vitals'])
gen_queries = set(['如何优化LCP', '生成式搜索引擎优化', 'GEO最佳实践'])

overlap = sc_queries & gen_queries  # 交集：两种引擎都覆盖
only_sc = sc_queries - gen_queries  # 仅传统搜索覆盖
only_gen = gen_queries - sc_queries # 仅生成引擎覆盖

print(f"双引擎覆盖查询: {overlap}")
print(f"需要加强GEO的查询: {only_sc}")
print(f"需要加强SEO的查询: {only_gen}")

import pandas as pd
import numpy as np
from scipy import stats

# 假设数据
data = {
    'query': ['如何优化LCP', 'SEO技术', 'GEO指南', '结构化数据'],
    'sc_avg_position': [1.5, 3.2, 5.1, 2.8],
    'gen_mention_position': [1, 2, 3, 1],
    'sc_clicks': [1200, 800, 300, 1500],
    'gen_mentions': [50, 30, 20, 45]
}
df = pd.DataFrame(data)

# 计算相关性
correlation = df['sc_avg_position'].corr(df['gen_mention_position'])
print(f"传统排名与生成引擎引用位置的相关性: {correlation}")

# 统计检验
slope, intercept, r_value, p_value, std_err = stats.linregress(
    df['sc_avg_position'], df['gen_mention_position']
)
print(f"p值: {p_value}")

import matplotlib.pyplot as plt
import seaborn as sns

# 假设时间序列数据
dates = pd.date_range('2024-01-01', periods=30, freq='D')
sc_data = np.random.randint(100, 200, size=30)
gen_data = np.random.randint(20, 50, size=30)

plt.figure(figsize=(12, 6))
plt.plot(dates, sc_data, label='Search Console Clicks', marker='o')
plt.plot(dates, gen_data, label='Gen Engine Mentions', marker='s')
plt.title('Search Console vs Gen Engine: Daily Performance')
plt.xlabel('Date')
plt.ylabel('Count')
plt.legend()
plt.grid(True)
plt.show()

plt.figure(figsize=(8, 6))
sns.scatterplot(data=df, x='sc_avg_position', y='gen_mention_position', 
                size='sc_clicks', hue='gen_mentions', sizes=(50, 500))
plt.title('Traditional Ranking vs Gen Engine Mention Position')
plt.xlabel('Average Position (Search Console)')
plt.ylabel('Mention Position (Gen Engine)')
plt.gca().invert_xaxis()  # 排名越小越好
plt.gca().invert_yaxis()  # 引用位置越小越好
plt.show()

class SearchEngineComparator:
    def __init__(self, sc_api_client, gen_engine_client):
        self.sc_client = sc_api_client
        self.gen_client = gen_engine_client
    
    def fetch_and_compare(self, queries, date_range):
        # 1. 获取 Search Console 数据
        sc_data = self.sc_client.get_performance_data(queries, date_range)
        
        # 2. 获取生成引擎引用数据
        gen_data = self.gen_client.get_mention_data(queries, date_range)
        
        # 3. 数据对齐
        merged = pd.merge(sc_data, gen_data, on=['query', 'date'], how='outer')
        
        # 4. 计算差异指标
        merged['coverage_gap'] = merged['gen_mentions'] - merged['sc_clicks']
        merged['rank_correlation'] = merged['sc_avg_position'].corr(merged['gen_mention_position'])
        
        # 5. 生成对比报告
        report = {
            'high_sc_low_gen': merged[merged['sc_clicks'] > 100][merged['gen_mentions'] < 10],
            'low_sc_high_gen': merged[merged['sc_clicks'] < 10][merged['gen_mentions'] > 20],
            'correlation': merged['rank_correlation'].mean()
        }
        
        return report