{"id":3889,"date":"2025-09-13T08:49:40","date_gmt":"2025-09-13T00:49:40","guid":{"rendered":"http:\/\/viplao.com\/?p=3889"},"modified":"2025-09-13T08:49:42","modified_gmt":"2025-09-13T00:49:42","slug":"%e3%80%90python%e5%ae%9e%e8%b7%b5%e6%a1%88%e4%be%8b%e3%80%91%e7%94%b5%e5%95%86%e5%b9%b3%e5%8f%b0%e6%95%b0%e6%8d%ae%e5%88%86%e6%9e%90%e5%92%8c%e6%8c%96%e6%8e%98-%e7%94%a8%e6%88%b7%e7%94%9f%e5%91%bd","status":"publish","type":"post","link":"http:\/\/viplao.com\/index.php\/2025\/09\/13\/%e3%80%90python%e5%ae%9e%e8%b7%b5%e6%a1%88%e4%be%8b%e3%80%91%e7%94%b5%e5%95%86%e5%b9%b3%e5%8f%b0%e6%95%b0%e6%8d%ae%e5%88%86%e6%9e%90%e5%92%8c%e6%8c%96%e6%8e%98-%e7%94%a8%e6%88%b7%e7%94%9f%e5%91%bd\/","title":{"rendered":"\u3010Python\u5b9e\u8df5\u6848\u4f8b\u3011\u7535\u5546\u5e73\u53f0\u6570\u636e\u5206\u6790\u548c\u6316\u6398 -\u7528\u6237\u751f\u547d\u5468\u671f\u7ba1\u7406"},"content":{"rendered":"\n<h3 class=\"wp-block-heading\">\u5f00\u53d1\u601d\u8def<\/h3>\n\n\n\n<ol>\n<li><strong>\u914d\u7f6e (<code># --- \u914d\u7f6e ---<\/code>)<\/strong>:\n<ul>\n<li><code>REPORT_PREFIX<\/code>: \u751f\u6210\u7684\u62a5\u544a\u548c\u56fe\u8868\u6587\u4ef6\u540d\u7684\u524d\u7f00\u3002<\/li>\n\n\n\n<li><code>NUM_USERS<\/code>,&nbsp;<code>START_DATE<\/code>,&nbsp;<code>END_DATE<\/code>: \u7528\u4e8e\u751f\u6210\u6a21\u62df\u6570\u636e\u7684\u53c2\u6570\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u6570\u636e\u751f\u6210 (<code>generate_sample_data<\/code>)<\/strong>:\n<ul>\n<li>\u6a21\u62df\u4e865000\u540d\u7528\u6237\u7684\u6570\u636e\uff0c\u5305\u62ec&nbsp;<code>user_id<\/code>,&nbsp;<code>signup_date<\/code>,&nbsp;<code>segment<\/code>&nbsp;(\u5206\u7fa4),&nbsp;<code>first_purchase_date<\/code>,&nbsp;<code>last_purchase_date<\/code>,&nbsp;<code>total_orders<\/code>,&nbsp;<code>total_spent<\/code>,&nbsp;<code>is_churned<\/code>&nbsp;(\u662f\u5426\u6d41\u5931)\u3002<\/li>\n\n\n\n<li>\u7528\u6237\u88ab\u5206\u4e3a\u4e09\u7c7b\uff1a<code>\u666e\u901a\u7528\u6237<\/code>,&nbsp;<code>\u9ad8\u4ef7\u503c\u7528\u6237<\/code>,&nbsp;<code>\u6d41\u5931\u98ce\u9669\u7528\u6237<\/code>\uff0c\u5e76\u6839\u636e\u7c7b\u522b\u5206\u914d\u4e0d\u540c\u7684\u884c\u4e3a\u7279\u5f81\uff08\u5982\u8d2d\u4e70\u9891\u7387\u3001\u6d88\u8d39\u91d1\u989d\uff09\u3002<\/li>\n\n\n\n<li>\u6d41\u5931\u5224\u5b9a\uff1a\u5728\u5206\u6790\u5468\u671f\u7ed3\u675f\u65f6\uff082024\u5e741\u67081\u65e5\uff09\uff0c\u5982\u679c\u7528\u6237\u8ddd\u79bb\u4e0a\u6b21\u8d2d\u4e70\u5df2\u8d85\u8fc790\u5929\uff0c\u5219\u6807\u8bb0\u4e3a\u6d41\u5931\u3002<\/li>\n\n\n\n<li>\u751f\u6210\u7684\u6570\u636e\u4fdd\u5b58\u4e3aCSV\u6587\u4ef6\uff0c\u65b9\u4fbf\u67e5\u770b\u548c\u540e\u7eed\u4f7f\u7528\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u5206\u6790\u51fd\u6570<\/strong>:\n<ul>\n<li><code>analyze_user_cohorts<\/code>:\n<ul>\n<li>\u8fdb\u884c&nbsp;<strong>Cohort \u7559\u5b58\u5206\u6790<\/strong>\u3002\u5c06\u7528\u6237\u6309\u6ce8\u518c\u5468\uff08Cohort Date\uff09\u5206\u7ec4\u3002<\/li>\n\n\n\n<li>\u8ffd\u8e2a\u6bcf\u4e2a\u7ec4\u7684\u7528\u6237\u5728\u540e\u7eed\u6bcf\u5468\u7684\u6d3b\u8dc3\u60c5\u51b5\uff08\u5373\u7559\u5b58\uff09\u3002<\/li>\n\n\n\n<li>\u8ba1\u7b97\u5e76\u8f93\u51fa\u4e00\u4e2a\u7559\u5b58\u7387\u77e9\u9635\uff0c\u663e\u793a\u6bcf\u4e2aCohort\u5728\u4e0d\u540c\u5468\u671f\u540e\u7684\u7559\u5b58\u767e\u5206\u6bd4\u3002<\/li>\n\n\n\n<li>\u4f7f\u7528&nbsp;<code>matplotlib<\/code>&nbsp;\u7ed8\u5236\u6700\u8fd1\u51e0\u4e2aCohort\u7684\u7559\u5b58\u7387\u66f2\u7ebf\u56fe\uff0c\u76f4\u89c2\u5c55\u793a\u7528\u6237\u6d41\u5931\u8d8b\u52bf\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><code>analyze_segments<\/code>:\n<ul>\n<li>\u6309\u7528\u6237\u5206\u7fa4 (<code>segment<\/code>) \u8ba1\u7b97\u5173\u952e\u6307\u6807\uff1a\u7528\u6237\u6570\u3001\u5e73\u5747\u8ba2\u5355\u6570\u3001\u5e73\u5747\u6d88\u8d39\u3001\u603b\u6d88\u8d39\u3001\u6d41\u5931\u7528\u6237\u6570\u3001\u6d41\u5931\u7387\u3002<\/li>\n\n\n\n<li>\u5206\u6790\u5404\u5206\u7fa4\u5728\u7528\u6237\u603b\u91cf\u548c\u6d88\u8d39\u603b\u989d\u4e2d\u7684\u5360\u6bd4\u3002<\/li>\n\n\n\n<li>\u4f7f\u7528&nbsp;<code>matplotlib<\/code>&nbsp;\u7ed8\u5236\u4e24\u4e2a\u997c\u56fe\uff0c\u5206\u522b\u5c55\u793a\u5404\u5206\u7fa4\u7684\u7528\u6237\u6570\u91cf\u5360\u6bd4\u548c\u6d88\u8d39\u91d1\u989d\u5360\u6bd4\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><code>analyze_ltv<\/code>:\n<ul>\n<li>\u8ba1\u7b97\u6bcf\u4e2a\u7528\u6237\u5206\u7fa4\u7684&nbsp;<strong>\u5e73\u5747\u751f\u547d\u5468\u671f\u4ef7\u503c (LTV)<\/strong>\uff0c\u5373\u8be5\u5206\u7fa4\u7528\u6237\u7684\u5e73\u5747\u603b\u6d88\u8d39\u3002<\/li>\n\n\n\n<li>\u4f7f\u7528&nbsp;<code>matplotlib<\/code>&nbsp;\u7ed8\u5236\u67f1\u72b6\u56fe\u5c55\u793a\u5404\u5206\u7fa4\u7684\u5e73\u5747LTV\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u62a5\u544a\u751f\u6210 (<code>generate_retention_strategies_report<\/code>)<\/strong>:\n<ul>\n<li>\u6c47\u603b\u6240\u6709\u5206\u6790\u7ed3\u679c\u3002<\/li>\n\n\n\n<li>\u8ba1\u7b97\u6574\u4f53\u6838\u5fc3\u6307\u6807\uff08\u603b\u7528\u6237\u6570\u3001\u603b\u6d41\u5931\u6570\u3001\u6574\u4f53\u6d41\u5931\u7387\uff09\u3002<\/li>\n\n\n\n<li>\u6253\u5370\u5404\u90e8\u5206\u7684\u5206\u6790\u6458\u8981\u548c\u6570\u636e\u8868\u683c\u3002<\/li>\n\n\n\n<li><strong>\u6838\u5fc3\u90e8\u5206<\/strong>: \u57fa\u4e8e\u6570\u636e\u5206\u6790\uff0c\u63d0\u51fa\u5177\u4f53\u7684\u3001\u53ef\u6267\u884c\u7684\u7528\u6237\u7559\u5b58\u7b56\u7565\u5efa\u8bae\u3002\u4f8b\u5982\uff1a\n<ul>\n<li>\u9488\u5bf9LTV\u6700\u9ad8\u7684\u5206\u7fa4\uff0c\u5efa\u8bae\u63d0\u4f9bVIP\u670d\u52a1\u3002<\/li>\n\n\n\n<li>\u9488\u5bf9\u6d41\u5931\u7387\u6700\u9ad8\u7684\u5206\u7fa4\uff0c\u5efa\u8bae\u8fdb\u884c\u633d\u7559\u6d3b\u52a8\u3002<\/li>\n\n\n\n<li>\u9488\u5bf9\u65b0\u7528\u6237\uff0c\u5efa\u8bae\u4f18\u5316\u5f15\u5bfc\u548c\u9996\u5355\u4f53\u9a8c\u3002<\/li>\n\n\n\n<li>\u63d0\u51fa\u901a\u7528\u7684\u7cbe\u7ec6\u5316\u8fd0\u8425\u5efa\u8bae\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li>\u5c06\u6240\u6709\u5185\u5bb9\u5199\u5165\u4e00\u4e2a&nbsp;<code>.txt<\/code>&nbsp;\u6587\u4ef6\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u4e3b\u51fd\u6570 (<code>main<\/code>)<\/strong>:\n<ul>\n<li>\u534f\u8c03\u6574\u4e2a\u6d41\u7a0b\uff1a\u751f\u6210\u6570\u636e\u3001\u8c03\u7528\u5206\u6790\u51fd\u6570\u3001\u751f\u6210\u62a5\u544a\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p><\/p>\n\n\n\n<p><\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>import pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom datetime import datetime, timedelta\nimport os\n\n# --- \u914d\u7f6e ---\n# \u62a5\u544a\u6587\u4ef6\u540d\u524d\u7f00\nREPORT_PREFIX = '\u7528\u6237\u751f\u547d\u5468\u671f\u4e0e\u7559\u5b58\u7b56\u7565\u62a5\u544a'\n# \u6a21\u62df\u6570\u636e\u53c2\u6570\nNUM_USERS = 5000\nSTART_DATE = datetime(2023, 1, 1)\nEND_DATE = datetime(2024, 1, 1)\n\n# --- \u6570\u636e\u751f\u6210\u51fd\u6570 ---\n\ndef generate_sample_data(n_users, start_date, end_date):\n    \"\"\"\u751f\u6210\u6a21\u62df\u7528\u6237\u751f\u547d\u5468\u671f\u6570\u636e\"\"\"\n    print(\"--- \u6b63\u5728\u751f\u6210\u6a21\u62df\u7528\u6237\u6570\u636e ---\")\n    np.random.seed(42) # \u4fdd\u8bc1\u7ed3\u679c\u53ef\u590d\u73b0\n\n    data = &#91;]\n    user_ids = range(1, n_users + 1)\n    \n    for user_id in user_ids:\n        # 1. \u6ce8\u518c\u65e5\u671f: \u5728\u6307\u5b9a\u65f6\u95f4\u8303\u56f4\u5185\u968f\u673a\n        signup_date = start_date + timedelta(days=np.random.randint(0, (end_date - start_date).days))\n        \n        # 2. \u7528\u6237\u5206\u7fa4 (\u7b80\u5316\u6a21\u62df)\n        # \u5047\u8bbe60%\u662f\u666e\u901a\u7528\u6237\uff0c30%\u662f\u9ad8\u4ef7\u503c\u7528\u6237\uff0c10%\u662f\u6d41\u5931\u98ce\u9669\u7528\u6237\n        segments = &#91;'\u666e\u901a\u7528\u6237', '\u9ad8\u4ef7\u503c\u7528\u6237', '\u6d41\u5931\u98ce\u9669\u7528\u6237']\n        segment_probs = &#91;0.6, 0.3, 0.1]\n        segment = np.random.choice(segments, p=segment_probs)\n        \n        # 3. \u9996\u6b21\u8d2d\u4e70\u65e5\u671f: \u6ce8\u518c\u540e0-30\u5929\u5185\n        days_to_first_purchase = np.random.randint(0, 31)\n        first_purchase_date = signup_date + timedelta(days=days_to_first_purchase)\n        if first_purchase_date &gt; end_date:\n            first_purchase_date = None # \u90e8\u5206\u7528\u6237\u6ce8\u518c\u540e\u672a\u8d2d\u4e70\n            \n        # 4. \u603b\u8d2d\u4e70\u6b21\u6570\u548c\u603b\u6d88\u8d39 (\u4e0e\u5206\u7fa4\u76f8\u5173)\n        if segment == '\u9ad8\u4ef7\u503c\u7528\u6237':\n            total_orders = np.random.poisson(15) + 1 # \u6cca\u677e\u5206\u5e03\uff0c\u786e\u4fdd\u81f3\u5c111\u6b21\n            total_spent = np.random.lognormal(10, 0.5) # \u5bf9\u6570\u6b63\u6001\u5206\u5e03\u6a21\u62df\u9ad8\u6d88\u8d39\n        elif segment == '\u666e\u901a\u7528\u6237':\n            total_orders = np.random.poisson(3) + 1\n            total_spent = np.random.lognormal(8, 0.8)\n        else: # \u6d41\u5931\u98ce\u9669\u7528\u6237\n            total_orders = np.random.poisson(1)\n            total_spent = np.random.lognormal(7, 1.0)\n            \n        # 5. \u6700\u540e\u8d2d\u4e70\u65e5\u671f (\u51b3\u5b9a\u662f\u5426\u5df2\u6d41\u5931)\n        # \u5047\u8bbe\u7528\u6237\u5e73\u5747\u6bcf\u6708\u8d2d\u4e70\u4e00\u6b21\uff0c\u6d41\u5931\u524d\u5e73\u5747\u6d3b\u8dc36\u4e2a\u6708\n        avg_days_between_purchases = 30\n        std_days = 10\n        days_active = np.random.normal(6 * avg_days_between_purchases, 2 * std_days)\n        days_active = max(days_active, avg_days_between_purchases) # \u81f3\u5c11\u6d3b\u8dc3\u4e00\u6b21\u8d2d\u4e70\u5468\u671f\n        \n        last_purchase_date = first_purchase_date + timedelta(days=days_active) if first_purchase_date else None\n        if last_purchase_date and last_purchase_date &gt; end_date:\n            last_purchase_date = end_date\n            \n        # 6. \u5224\u5b9a\u662f\u5426\u6d41\u5931 (\u5728\u5206\u6790\u5468\u671f\u7ed3\u675f\u65f6\uff0c\u8ddd\u79bb\u4e0a\u6b21\u8d2d\u4e70\u8d85\u8fc790\u5929)\n        is_churned = False\n        if last_purchase_date:\n            days_since_last_purchase = (end_date - last_purchase_date).days\n            is_churned = days_since_last_purchase &gt; 90\n            \n        data.append({\n            'user_id': user_id,\n            'signup_date': signup_date,\n            'segment': segment,\n            'first_purchase_date': first_purchase_date,\n            'last_purchase_date': last_purchase_date,\n            'total_orders': total_orders,\n            'total_spent': round(total_spent, 2),\n            'is_churned': is_churned\n        })\n        \n    df = pd.DataFrame(data)\n    # \u4fee\u6b63\u672a\u8d2d\u4e70\u7528\u6237\u7684\u5b57\u6bb5\n    df.loc&#91;df&#91;'first_purchase_date'].isnull(), &#91;'last_purchase_date', 'total_orders', 'total_spent']] = &#91;None, 0, 0]\n    \n    csv_filename = f'{REPORT_PREFIX}_\u6a21\u62df\u6570\u636e.csv'\n    df.to_csv(csv_filename, index=False, encoding='utf-8-sig') # utf-8-sig for Excel compatibility\n    print(f\"\u6a21\u62df\u6570\u636e\u5df2\u751f\u6210\u5e76\u4fdd\u5b58\u81f3: {csv_filename}\")\n    return df\n\n# --- \u5206\u6790\u51fd\u6570 ---\n\ndef analyze_user_cohorts(df, analysis_end_date):\n    \"\"\"\u5206\u6790\u7528\u6237\u7559\u5b58\u7387 (\u6309\u6ce8\u518c\u5468 cohorts)\"\"\"\n    print(\"\\n--- \u6b63\u5728\u5206\u6790\u7528\u6237\u7559\u5b58\u7387 (Cohort Analysis) ---\")\n    \n    # \u786e\u4fdd\u65e5\u671f\u5217\u662fdatetime\u7c7b\u578b\n    df&#91;'signup_date'] = pd.to_datetime(df&#91;'signup_date'])\n    df&#91;'last_purchase_date'] = pd.to_datetime(df&#91;'last_purchase_date'])\n\n    # 1. \u5b9a\u4e49 Cohort (\u4ee5\u7528\u6237\u6ce8\u518c\u5468\u7684\u7b2c\u4e00\u5929\u4e3a Cohort Date)\n    df&#91;'cohort_date'] = df&#91;'signup_date'].dt.to_period('W').dt.start_time\n    \n    # 2. \u8ba1\u7b97\u6bcf\u4e2a\u7528\u6237\u6700\u540e\u4e00\u6b21\u6d3b\u8dc3\u7684\u65f6\u671f\n    # \u5982\u679c\u5df2\u6d41\u5931\uff0c\u5219\u6d3b\u8dc3\u622a\u6b62\u5230\u6d41\u5931\u524d\uff1b\u5982\u679c\u672a\u6d41\u5931\uff0c\u5219\u622a\u6b62\u5230\u5206\u6790\u671f\u672b\n    df&#91;'active_end_date'] = np.where(df&#91;'is_churned'], df&#91;'last_purchase_date'], analysis_end_date)\n    df&#91;'active_end_period'] = df&#91;'active_end_date'].dt.to_period('W').dt.start_time\n    \n    # 3. \u8ba1\u7b97\u7528\u6237\u6240\u5c5e\u7684cohort\u548c\u6d3b\u8dc3\u7684cohort\n    cohort_groups = df.groupby(df&#91;'cohort_date'])\n    \n    # 4. \u6784\u5efa\u7559\u5b58\u7387\u77e9\u9635\n    cohort_sizes = cohort_groups&#91;'user_id'].count() # \u6bcf\u4e2acohort\u7684\u521d\u59cb\u7528\u6237\u6570\n    \n    # \u4e3a\u6bcf\u4e2a\u7528\u6237\u751f\u6210\u5176\u6d3b\u8dc3\u671f\u5185\u7684\u6240\u6709\u5468\n    user_cohort_periods = &#91;]\n    for _, user in df.iterrows():\n        if pd.notnull(user&#91;'cohort_date']) and pd.notnull(user&#91;'active_end_period']):\n            periods_active = pd.date_range(user&#91;'cohort_date'], user&#91;'active_end_period'], freq='W-MON')\n            for period in periods_active:\n                user_cohort_periods.append({'cohort_date': user&#91;'cohort_date'], 'active_period': period})\n\n    df_active_periods = pd.DataFrame(user_cohort_periods)\n    \n    # \u8ba1\u7b97\u6bcf\u4e2acohort\u5728\u4e0d\u540c\u5468\u671f\u7684\u6d3b\u8dc3\u7528\u6237\u6570\n    retention_counts = df_active_periods.groupby(&#91;'cohort_date', 'active_period']).size().reset_index(name='active_users')\n    \n    # \u5c06\u5468\u671f\u8f6c\u6362\u4e3a\u76f8\u5bf9\u4e8ecohort\u7684\u6708\u4efd\u7d22\u5f15 (\u7b80\u5316\u4e3a\u5468)\n    retention_counts&#91;'period_number'] = ((retention_counts&#91;'active_period'] - retention_counts&#91;'cohort_date']).dt.days \/ 7).astype(int)\n    \n    # \u900f\u89c6\u8868\uff1a\u884c\u662fcohort\uff0c\u5217\u662f\u5468\u671f\uff0c\u503c\u662f\u6d3b\u8dc3\u7528\u6237\u6570\n    retention_pivot = retention_counts.pivot_table(index='cohort_date', columns='period_number', values='active_users', fill_value=0)\n    \n    # \u8ba1\u7b97\u7559\u5b58\u7387 (%) \n    cohort_sizes_aligned = cohort_sizes.reindex(retention_pivot.index, fill_value=0)\n    retention_matrix = retention_pivot.divide(cohort_sizes_aligned, axis=0) * 100\n\n    print(\"\u7528\u6237\u5468\u7559\u5b58\u7387\u77e9\u9635 (\u90e8\u5206\u6570\u636e):\")\n    print(retention_matrix.head(10).round(1)) # \u663e\u793a\u524d10\u4e2acohort\n    \n    # \u7ed8\u5236\u7559\u5b58\u7387\u66f2\u7ebf (\u9009\u53d6\u51e0\u4e2a\u6700\u8fd1\u7684cohort)\n    plt.figure(figsize=(12, 6))\n    recent_cohorts = retention_matrix.tail(5).index # \u6700\u8fd15\u4e2acohort\n    for cohort_date in recent_cohorts:\n        cohort_data = retention_matrix.loc&#91;cohort_date].dropna()\n        periods = cohort_data.index\n        rates = cohort_data.values\n        plt.plot(periods, rates, marker='o', label=f'Cohort: {cohort_date.strftime(\"%Y-%m-%d\")}')\n        \n    plt.xlabel('\u5468\u6570 (\u76f8\u5bf9\u4e8e\u6ce8\u518c\u5468)')\n    plt.ylabel('\u7559\u5b58\u7387 (%)')\n    plt.title('\u7528\u6237\u6309\u5468 Cohort \u7559\u5b58\u7387\u8d8b\u52bf')\n    plt.legend()\n    plt.grid(True)\n    cohort_chart_path = f'{REPORT_PREFIX}_\u7559\u5b58\u7387\u66f2\u7ebf.png'\n    plt.savefig(cohort_chart_path)\n    plt.close()\n    print(f\"Cohort\u7559\u5b58\u7387\u56fe\u8868\u5df2\u4fdd\u5b58\u81f3: {cohort_chart_path}\")\n    \n    return retention_matrix, cohort_chart_path\n\ndef analyze_segments(df):\n    \"\"\"\u5206\u6790\u4e0d\u540c\u7528\u6237\u5206\u7fa4\u7684\u7279\u5f81\"\"\"\n    print(\"\\n--- \u6b63\u5728\u5206\u6790\u7528\u6237\u5206\u7fa4\u7279\u5f81 ---\")\n    \n    # \u6309\u5206\u7fa4\u8ba1\u7b97\u5173\u952e\u6307\u6807\n    segment_analysis = df.groupby('segment').agg(\n        user_count=('user_id', 'count'),\n        avg_orders=('total_orders', 'mean'),\n        avg_spent=('total_spent', 'mean'),\n        total_spent=('total_spent', 'sum'),\n        churned_count=('is_churned', 'sum')\n    ).reset_index()\n    \n    segment_analysis&#91;'churn_rate'] = (segment_analysis&#91;'churned_count'] \/ segment_analysis&#91;'user_count']) * 100\n    segment_analysis&#91;'avg_spent'] = segment_analysis&#91;'avg_spent'].round(2)\n    segment_analysis&#91;'avg_orders'] = segment_analysis&#91;'avg_orders'].round(2)\n    segment_analysis&#91;'churn_rate'] = segment_analysis&#91;'churn_rate'].round(2)\n    segment_analysis&#91;'pct_total_spent'] = (segment_analysis&#91;'total_spent'] \/ segment_analysis&#91;'total_spent'].sum()) * 100\n    segment_analysis&#91;'pct_total_spent'] = segment_analysis&#91;'pct_total_spent'].round(2)\n    \n    # \u91cd\u65b0\u6392\u5e8f\u5217\n    segment_analysis = segment_analysis&#91;&#91;\n        'segment', 'user_count', 'avg_orders', 'avg_spent', \n        'total_spent', 'pct_total_spent', 'churned_count', 'churn_rate'\n    ]]\n    \n    print(\"\u5404\u7528\u6237\u5206\u7fa4\u6838\u5fc3\u6307\u6807:\")\n    print(segment_analysis.to_string(index=False))\n    \n    # \u7ed8\u5236\u5206\u7fa4\u7528\u6237\u6570\u548c\u8d21\u732e\u989d\u997c\u56fe\n    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 7))\n    \n    # \u7528\u6237\u6570\u5206\u5e03\n    ax1.pie(segment_analysis&#91;'user_count'], labels=segment_analysis&#91;'segment'], autopct='%1.1f%%', startangle=140)\n    ax1.set_title('\u5404\u5206\u7fa4\u7528\u6237\u6570\u91cf\u5360\u6bd4')\n    \n    # \u6d88\u8d39\u8d21\u732e\u5206\u5e03\n    ax2.pie(segment_analysis&#91;'total_spent'], labels=segment_analysis&#91;'segment'], autopct='%1.1f%%', startangle=140)\n    ax2.set_title('\u5404\u5206\u7fa4\u6d88\u8d39\u91d1\u989d\u5360\u6bd4')\n    \n    plt.tight_layout()\n    segments_chart_path = f'{REPORT_PREFIX}_\u7528\u6237\u5206\u7fa4\u5206\u6790.png'\n    plt.savefig(segments_chart_path)\n    plt.close()\n    print(f\"\u7528\u6237\u5206\u7fa4\u5206\u6790\u56fe\u8868\u5df2\u4fdd\u5b58\u81f3: {segments_chart_path}\")\n    \n    return segment_analysis, segments_chart_path\n\ndef analyze_ltv(df):\n    \"\"\"\u5206\u6790\u7528\u6237\u751f\u547d\u5468\u671f\u4ef7\u503c (\u7b80\u5316\u8ba1\u7b97)\"\"\"\n    print(\"\\n--- \u6b63\u5728\u5206\u6790\u7528\u6237\u751f\u547d\u5468\u671f\u4ef7\u503c (LTV) ---\")\n    \n    # \u7b80\u5316LTV\u8ba1\u7b97\uff1a\u603b\u6d88\u8d39 \/ \u603b\u7528\u6237\u6570 (\u5e73\u5747LTV)\n    # \u6216\u6309\u5206\u7fa4\u8ba1\u7b97\u5e73\u5747LTV\n    ltv_analysis = df.groupby('segment').agg(\n        user_count=('user_id', 'count'),\n        total_spent=('total_spent', 'sum')\n    ).reset_index()\n    ltv_analysis&#91;'avg_ltv'] = (ltv_analysis&#91;'total_spent'] \/ ltv_analysis&#91;'user_count']).round(2)\n    \n    print(\"\u5404\u5206\u7fa4\u5e73\u5747\u751f\u547d\u5468\u671f\u4ef7\u503c (LTV):\")\n    print(ltv_analysis&#91;&#91;'segment', 'avg_ltv']].to_string(index=False))\n    \n    # \u7ed8\u5236LTV\u67f1\u72b6\u56fe\n    plt.figure(figsize=(10, 6))\n    bars = plt.bar(ltv_analysis&#91;'segment'], ltv_analysis&#91;'avg_ltv'], color=&#91;'skyblue', 'gold', 'lightcoral'])\n    plt.xlabel('\u7528\u6237\u5206\u7fa4')\n    plt.ylabel('\u5e73\u5747\u751f\u547d\u5468\u671f\u4ef7\u503c (\u5143)')\n    plt.title('\u5404\u5206\u7fa4\u5e73\u5747\u751f\u547d\u5468\u671f\u4ef7\u503c (LTV)')\n    \n     # \u5728\u67f1\u72b6\u56fe\u4e0a\u6dfb\u52a0\u6570\u503c\u6807\u7b7e\n    for bar in bars:\n        yval = bar.get_height()\n        plt.text(bar.get_x() + bar.get_width()\/2.0, yval, round(yval, 2), ha='center', va='bottom')\n\n    plt.tight_layout()\n    ltv_chart_path = f'{REPORT_PREFIX}_LTV\u5206\u6790.png'\n    plt.savefig(ltv_chart_path)\n    plt.close()\n    print(f\"LTV\u5206\u6790\u56fe\u8868\u5df2\u4fdd\u5b58\u81f3: {ltv_chart_path}\")\n    \n    return ltv_analysis, ltv_chart_path\n\n# --- \u62a5\u544a\u751f\u6210\u51fd\u6570 ---\n\ndef generate_retention_strategies_report(df, retention_matrix, cohort_chart, segment_df, segments_chart, ltv_df, ltv_chart):\n    \"\"\"\u751f\u6210\u7559\u5b58\u7b56\u7565\u62a5\u544a\"\"\"\n    print(\"\\n--- \u6b63\u5728\u751f\u6210\u7559\u5b58\u7b56\u7565\u62a5\u544a ---\")\n    from datetime import datetime\n    report_filename = f\"{REPORT_PREFIX}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.txt\"\n    \n    total_users = len(df)\n    total_churned = df&#91;'is_churned'].sum()\n    overall_churn_rate = (total_churned \/ total_users) * 100 if total_users &gt; 0 else 0\n    \n    with open(report_filename, 'w', encoding='utf-8') as f:\n        f.write(\"=\" * 50 + \"\\n\")\n        f.write(\"        \u7535\u5546\u5e73\u53f0\u7528\u6237\u751f\u547d\u5468\u671f\u4e0e\u7559\u5b58\u7b56\u7565\u62a5\u544a\\n\")\n        f.write(f\"        \u751f\u6210\u65f6\u95f4: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\\n\")\n        f.write(\"=\" * 50 + \"\\n\\n\")\n\n        f.write(\"--- 1. \u6838\u5fc3\u6307\u6807\u6982\u89c8 ---\\n\")\n        f.write(f\"\u603b\u7528\u6237\u6570: {total_users}\\n\")\n        f.write(f\"\u603b\u6d41\u5931\u7528\u6237\u6570: {total_churned}\\n\")\n        f.write(f\"\u6574\u4f53\u6d41\u5931\u7387: {overall_churn_rate:.2f}%\\n\\n\")\n\n        f.write(\"--- 2. \u7528\u6237\u7559\u5b58\u7387\u5206\u6790 (Cohort Analysis) ---\\n\")\n        f.write(\"\u8bf4\u660e\uff1a\u8ffd\u8e2a\u540c\u4e00\u65f6\u95f4\u6ce8\u518c\u7684\u7528\u6237\u7fa4\u968f\u65f6\u95f4\u7684\u7559\u5b58\u60c5\u51b5\u3002\\n\")\n        f.write(\"\u89e3\u8bfb\uff1a\u56fe\u8868\u663e\u793a\u4e86\u6700\u8fd1\u51e0\u4e2a\u7528\u6237\u7fa4\u7684\u6309\u5468\u7559\u5b58\u7387\u3002\u7406\u60f3\u60c5\u51b5\u4e0b\uff0c\u66f2\u7ebf\u5e94\u7f13\u6162\u4e0b\u964d\u3002\\n\")\n        f.write(\"      \u5feb\u901f\u4e0b\u964d\u53ef\u80fd\u8868\u660e\u65b0\u7528\u6237\u4f53\u9a8c\u6216\u65e9\u671f\u4ef7\u503c\u4f20\u9012\u5b58\u5728\u95ee\u9898\u3002\\n\")\n        f.write(f\"\u56fe\u8868: {cohort_chart}\\n\")\n        f.write(\"(\u8bf7\u67e5\u770bCSV\u6587\u4ef6\u83b7\u53d6\u5b8c\u6574\u7684\u7559\u5b58\u7387\u77e9\u9635\u6570\u636e)\\n\\n\")\n\n        f.write(\"--- 3. \u7528\u6237\u5206\u7fa4\u7279\u5f81\u5206\u6790 ---\\n\")\n        f.write(\"\u8bf4\u660e\uff1a\u5c06\u7528\u6237\u5206\u4e3a\u4e0d\u540c\u7fa4\u4f53\uff0c\u5206\u6790\u5176\u884c\u4e3a\u5dee\u5f02\u3002\\n\")\n        f.write(segment_df.to_string(index=False))\n        f.write(f\"\\n\u56fe\u8868: {segments_chart}\\n\\n\")\n\n        f.write(\"--- 4. \u751f\u547d\u5468\u671f\u4ef7\u503c (LTV) \u5206\u6790 ---\\n\")\n        f.write(\"\u8bf4\u660e\uff1a\u8861\u91cf\u7528\u6237\u5728\u751f\u547d\u5468\u671f\u5185\u4e3a\u4f01\u4e1a\u521b\u9020\u7684\u5e73\u5747\u6536\u5165\u3002\\n\")\n        f.write(ltv_df&#91;&#91;'segment', 'avg_ltv']].to_string(index=False))\n        f.write(f\"\\n\u56fe\u8868: {ltv_chart}\\n\\n\")\n\n        f.write(\"--- 5. \u7559\u5b58\u7b56\u7565\u5efa\u8bae ---\\n\")\n        f.write(\"\u57fa\u4e8e\u4ee5\u4e0a\u5206\u6790\uff0c\u63d0\u51fa\u4ee5\u4e0b\u9488\u5bf9\u6027\u7684\u7528\u6237\u7559\u5b58\u7b56\u7565\uff1a\\n\\n\")\n        \n        high_value_seg = ltv_df.loc&#91;ltv_df&#91;'avg_ltv'].idxmax(), 'segment']\n        high_churn_seg_row = segment_df.loc&#91;segment_df&#91;'churn_rate'].idxmax()]\n        high_churn_seg = high_churn_seg_row&#91;'segment']\n        high_churn_rate = high_churn_seg_row&#91;'churn_rate']\n        \n        f.write(f\"1. \u91cd\u70b9\u7ef4\u62a4\u9ad8\u4ef7\u503c\u7528\u6237 ({high_value_seg}):\\n\")\n        f.write(f\"   - \u7279\u5f81: \u5e73\u5747LTV\u6700\u9ad8\u3002\\n\")\n        f.write(f\"   - \u7b56\u7565: \u63d0\u4f9bVIP\u670d\u52a1\u3001\u4e13\u5c5e\u4f18\u60e0\u3001\u4f18\u5148\u5ba2\u670d\u3001\u65b0\u54c1\u4f53\u9a8c\u6743\u7b49\uff0c\u63d0\u5347\u5176\u5fe0\u8bda\u5ea6\u548c\u6ee1\u610f\u5ea6\u3002\\n\\n\")\n        \n        f.write(f\"2. \u633d\u6551\u6d41\u5931\u98ce\u9669\u7528\u6237 ({high_churn_seg}):\\n\")\n        f.write(f\"   - \u7279\u5f81: \u6d41\u5931\u7387\u6700\u9ad8 ({high_churn_rate:.2f}%)\u3002\\n\")\n        f.write(f\"   - \u7b56\u7565: \u5206\u6790\u5176\u6d41\u5931\u539f\u56e0\uff0c\u901a\u8fc7\u90ae\u4ef6\/SMS\u63a8\u9001\u4e2a\u6027\u5316\u4f18\u60e0\u5238\u3001\u53ec\u56de\u6d3b\u52a8\uff0c\u6216\u8fdb\u884c\u7528\u6237\u56de\u8bbf\u4e86\u89e3\u9700\u6c42\u3002\\n\\n\")\n        \n        f.write(\"3. \u63d0\u5347\u65b0\u7528\u6237\u4ef7\u503c:\\n\")\n        f.write(\"   - \u7279\u5f81: \u65b0\u7528\u6237\uff08\u5982Cohort\u5206\u6790\u6240\u793a\uff09\u65e9\u671f\u6d41\u5931\u7387\u662f\u5173\u952e\u3002\\n\")\n        f.write(\"   - \u7b56\u7565: \u4f18\u5316\u65b0\u7528\u6237\u5f15\u5bfc\u6d41\u7a0b\uff0c\u63d0\u4f9b\u9996\u5355\u4f18\u60e0\uff0c\u8bbe\u7f6e\u65b0\u624b\u4efb\u52a1\u5956\u52b1\uff0c\u786e\u4fdd\u9996\u6b21\u8d2d\u4e70\u4f53\u9a8c\u987a\u7545\u3002\\n\\n\")\n        \n        f.write(\"4. \u901a\u7528\u7cbe\u7ec6\u5316\u8fd0\u8425:\\n\")\n        f.write(\"   - \u7b56\u7565: \u5229\u7528\u7528\u6237\u884c\u4e3a\u6570\u636e\u8fdb\u884c\u4e2a\u6027\u5316\u63a8\u8350\uff0c\u5b9e\u65bd\u751f\u547d\u5468\u671f\u81ea\u52a8\u5316\u8425\u9500\uff08\u5982\u6ce8\u518c\u540e\u3001\u9996\u8d2d\u540e\u3001\u6c89\u9ed8\u671f\u81ea\u52a8\u89e6\u53d1\u90ae\u4ef6\/SMS\uff09\u3002\\n\")\n        f.write(\"          \u5efa\u7acb\u79ef\u5206\u4f53\u7cfb\u3001\u4f1a\u5458\u7b49\u7ea7\u5236\u5ea6\uff0c\u589e\u52a0\u7528\u6237\u7c98\u6027\u3002\\n\\n\")\n\n        f.write(\"=\" * 50 + \"\\n\")\n        f.write(\"                    \u62a5\u544a\u7ed3\u675f\\n\")\n        f.write(\"=\" * 50 + \"\\n\")\n\n    print(f\"\u7559\u5b58\u7b56\u7565\u62a5\u544a\u5df2\u751f\u6210: {report_filename}\")\n\n# --- \u4e3b\u51fd\u6570 ---\n\ndef main():\n    \"\"\"\u4e3b\u51fd\u6570\"\"\"\n    # 1. \u751f\u6210\u6216\u52a0\u8f7d\u6570\u636e\n    # \u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u8fd9\u91cc\u4f1a\u662f\u52a0\u8f7d\u771f\u5b9e\u6570\u636e\u5e93\u6570\u636e\u7684\u903b\u8f91\n    df_users = generate_sample_data(NUM_USERS, START_DATE, END_DATE)\n    \n    # 2. \u6267\u884c\u5206\u6790\n    retention_matrix, cohort_chart = analyze_user_cohorts(df_users, END_DATE)\n    segment_df, segments_chart = analyze_segments(df_users)\n    ltv_df, ltv_chart = analyze_ltv(df_users)\n    \n    # 3. \u751f\u6210\u62a5\u544a\n    generate_retention_strategies_report(\n        df_users, retention_matrix, cohort_chart, \n        segment_df, segments_chart, \n        ltv_df, ltv_chart\n    )\n    \n    print(\"\\n\u5206\u6790\u5b8c\u6210\u3002\")\n\nif __name__ == \"__main__\":\n    main()\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"<p>\u5f00\u53d1\u601d\u8def<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[28],"views":341,"_links":{"self":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/3889"}],"collection":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/comments?post=3889"}],"version-history":[{"count":2,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/3889\/revisions"}],"predecessor-version":[{"id":3909,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/3889\/revisions\/3909"}],"wp:attachment":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/media?parent=3889"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/categories?post=3889"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/tags?post=3889"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}