pip install pandas numpy matplotlib seaborn jupyter

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# 設(shè)置中文字體
plt.rcParams['font.sans-serif'] = ['SimHei']  # 用來正常顯示中文標簽
plt.rcParams['axes.unicode_minus'] = False  # 用來正常顯示負號

# 加載數(shù)據(jù)
df = pd.read_csv('sales_data.csv')

# 查看數(shù)據(jù)基本信息
print(df.info())

# 查看數(shù)據(jù)前幾行
print(df.head())

# 檢查缺失值
print(df.isnull().sum())

# 處理缺失值
df = df.dropna()

# 轉(zhuǎn)換數(shù)據(jù)類型
df['order_date'] = pd.to_datetime(df['order_date'])

# 添加新列
df['year'] = df['order_date'].dt.year
df['month'] = df['order_date'].dt.month
df['day'] = df['order_date'].dt.day

# 統(tǒng)計描述
print(df.describe())

# 銷售趨勢分析
sales_by_date = df.groupby('order_date')['total_amount'].sum()

plt.figure(figsize=(12, 6))
sales_by_date.plot()
plt.title('每日銷售趨勢')
plt.xlabel('日期')
plt.ylabel('銷售額')
plt.show()

# 月度銷售趨勢
sales_by_month = df.groupby(['year', 'month'])['total_amount'].sum().reset_index()
sales_by_month['date'] = pd.to_datetime(sales_by_month[['year', 'month']].assign(day=1))

plt.figure(figsize=(12, 6))
sns.lineplot(x='date', y='total_amount', data=sales_by_month)
plt.title('月度銷售趨勢')
plt.xlabel('日期')
plt.ylabel('銷售額')
plt.show()

# 產(chǎn)品類別銷售分析
sales_by_category = df.groupby('category')['total_amount'].sum().sort_values(ascending=False)

plt.figure(figsize=(12, 6))
sales_by_category.plot(kind='bar')
plt.title('各產(chǎn)品類別銷售額')
plt.xlabel('產(chǎn)品類別')
plt.ylabel('銷售額')
plt.xticks(rotation=45)
plt.show()

# 產(chǎn)品銷售分析
top_10_products = df.groupby('product_name')['total_amount'].sum().sort_values(ascending=False).head(10)

plt.figure(figsize=(12, 6))
top_10_products.plot(kind='bar')
plt.title('銷售額前10的產(chǎn)品')
plt.xlabel('產(chǎn)品名稱')
plt.ylabel('銷售額')
plt.xticks(rotation=45)
plt.show()

# 支付方式分析
payment_method_count = df.groupby('payment_method')['order_id'].count()

plt.figure(figsize=(10, 6))
plt.pie(payment_method_count, labels=payment_method_count.index, autopct='%1.1f%%')
plt.title('支付方式分布')
plt.show()

# 客戶購買行為分析
customer_purchase = df.groupby('customer_id')['total_amount'].agg(['count', 'sum', 'mean']).rename(columns={'count': '購買次數(shù)', 'sum': '總消費', 'mean': '平均消費'})

plt.figure(figsize=(12, 6))
sns.histplot(customer_purchase['購買次數(shù)'], bins=20)
plt.title('客戶購買次數(shù)分布')
plt.xlabel('購買次數(shù)')
plt.ylabel('客戶數(shù)')
plt.show()

plt.figure(figsize=(12, 6))
sns.histplot(customer_purchase['總消費'], bins=20)
plt.title('客戶總消費分布')
plt.xlabel('總消費')
plt.ylabel('客戶數(shù)')
plt.show()

# 客戶價值分析
# 計算客戶生命周期價值 (CLV)
# 這里簡化計算，使用總消費作為CLV
customer_clv = df.groupby('customer_id')['total_amount'].sum().sort_values(ascending=False)

# 客戶分層
top_20_percent = int(len(customer_clv) * 0.2)
top_customers = customer_clv.head(top_20_percent)
print(f"前20%客戶數(shù)量: {top_20_percent}")
print(f"前20%客戶消費占比: {top_customers.sum() / customer_clv.sum():.2f}")

# 產(chǎn)品關(guān)聯(lián)分析
# 計算產(chǎn)品之間的關(guān)聯(lián)度
from itertools import combinations
from collections import defaultdict

# 構(gòu)建購物籃
baskets = df.groupby('order_id')['product_name'].apply(list).tolist()

# 計算產(chǎn)品對的出現(xiàn)次數(shù)
product_pairs = defaultdict(int)
for basket in baskets:
    if len(basket) >= 2:
        for pair in combinations(set(basket), 2):
            product_pairs[tuple(sorted(pair))] += 1

# 轉(zhuǎn)換為DataFrame
product_pairs_df = pd.DataFrame.from_dict(product_pairs, orient='index', columns=['count']).reset_index()
product_pairs_df[['product1', 'product2']] = pd.DataFrame(product_pairs_df['index'].tolist(), index=product_pairs_df.index)
product_pairs_df = product_pairs_df.drop('index', axis=1)
product_pairs_df = product_pairs_df.sort_values('count', ascending=False)

print("產(chǎn)品關(guān)聯(lián)度前10:")
print(product_pairs_df.head(10))

# 銷售預(yù)測
# 使用移動平均法進行銷售預(yù)測
from statsmodels.tsa.holtwinters import SimpleExpSmoothing

# 準備數(shù)據(jù)
sales_series = sales_by_date

# 擬合模型
model = SimpleExpSmoothing(sales_series).fit(smoothing_level=0.6, optimized=False)

# 預(yù)測未來7天
forecast = model.forecast(7)

# 可視化預(yù)測結(jié)果
plt.figure(figsize=(12, 6))
plt.plot(sales_series.index, sales_series.values, label='實際銷售額')
plt.plot(forecast.index, forecast.values, label='預(yù)測銷售額', linestyle='--')
plt.title('銷售額預(yù)測')
plt.xlabel('日期')
plt.ylabel('銷售額')
plt.legend()
plt.show()

# 生成綜合分析報告
import io
from PIL import Image

# 創(chuàng)建一個HTML報告
report_html = """
<!DOCTYPE html>
<html lang="zh-CN">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>電商銷售數(shù)據(jù)分析報告</title>
    <style>
        body { font-family: Arial, sans-serif; margin: 20px; }
        h1 { color: #333; }
        h2 { color: #555; }
        .section { margin-bottom: 30px; }
        .chart { margin: 20px 0; }
        table { border-collapse: collapse; width: 100%; margin: 20px 0; }
        th, td { border: 1px solid #ddd; padding: 8px; text-align: left; }
        th { background-color: #f2f2f2; }
    </style>
</head>
<body>
    <h1>電商銷售數(shù)據(jù)分析報告</h1>
    
    <div class="section">
        <h2>1. 數(shù)據(jù)概覽</h2>
        <p>本次分析基于電商平臺的銷售數(shù)據(jù)，包含訂單信息、客戶信息、產(chǎn)品信息等。</p>
        <p>數(shù)據(jù)量：{total_orders} 條訂單記錄</p>
        <p>時間范圍：{start_date} 至 {end_date}</p>
        <p>總銷售額：￥{total_sales:.2f}</p>
        <p>客戶數(shù)量：{total_customers} 位</p>
        <p>產(chǎn)品種類：{total_products} 種</p>
    </div>
    
    <div class="section">
        <h2>2. 銷售趨勢分析</h2>
        <p>從銷售趨勢圖可以看出，銷售額呈現(xiàn)[上升/下降/穩(wěn)定]趨勢。</p>
        <div class="chart">
            <img src="data:image/png;base64,{sales_trend_chart}" alt="銷售趨勢圖">
        </div>
    </div>
    
    <div class="section">
        <h2>3. 產(chǎn)品分析</h2>
        <p>銷售額最高的產(chǎn)品類別是 {top_category}，占總銷售額的 {top_category_percent:.2f}%。</p>
        <p>銷售額最高的產(chǎn)品是 {top_product}，銷售額為 ￥{top_product_sales:.2f}。</p>
        <div class="chart">
            <img src="data:image/png;base64,{category_chart}" alt="產(chǎn)品類別銷售圖">
        </div>
        <div class="chart">
            <img src="data:image/png;base64,{product_chart}" alt="產(chǎn)品銷售圖">
        </div>
    </div>
    
    <div class="section">
        <h2>4. 客戶分析</h2>
        <p>前20%的客戶貢獻了 {top_customer_percent:.2f}% 的銷售額，體現(xiàn)了帕累托法則（80/20法則）。</p>
        <div class="chart">
            <img src="data:image/png;base64,{customer_chart}" alt="客戶消費分布圖">
        </div>
    </div>
    
    <div class="section">
        <h2>5. 支付方式分析</h2>
        <p>最常用的支付方式是 {top_payment_method}，占比 {top_payment_percent:.2f}%。</p>
        <div class="chart">
            <img src="data:image/png;base64,{payment_chart}" alt="支付方式分布圖">
        </div>
    </div>
    
    <div class="section">
        <h2>6. 產(chǎn)品關(guān)聯(lián)分析</h2>
        <p>最常一起購買的產(chǎn)品組合是：</p>
        <table>
            <tr>
                <th>產(chǎn)品1</th>
                <th>產(chǎn)品2</th>
                <th>共同購買次數(shù)</th>
            </tr>
            {product_pairs_table}
        </table>
    </div>
    
    <div class="section">
        <h2>7. 銷售預(yù)測</h2>
        <p>基于歷史數(shù)據(jù)，預(yù)測未來7天的銷售額：</p>
        <div class="chart">
            <img src="data:image/png;base64,{forecast_chart}" alt="銷售預(yù)測圖">
        </div>
    </div>
    
    <div class="section">
        <h2>8. 結(jié)論與建議</h2>
        <p>1. <strong>銷售策略</strong>：重點關(guān)注銷售額高的產(chǎn)品類別和產(chǎn)品，加大促銷力度。</p>
        <p>2. <strong>客戶策略</strong>：針對高價值客戶，提供個性化服務(wù)和專屬優(yōu)惠，提高客戶忠誠度。</p>
        <p>3. <strong>產(chǎn)品策略</strong>：基于產(chǎn)品關(guān)聯(lián)分析，優(yōu)化產(chǎn)品布局和推薦系統(tǒng)，提高交叉銷售。</p>
        <p>4. <strong>庫存管理</strong>：根據(jù)銷售趨勢和預(yù)測，合理安排庫存，避免積壓和缺貨。</p>
        <p>5. <strong>支付方式</strong>：優(yōu)化支付流程，支持更多便捷的支付方式，提高轉(zhuǎn)化率。</p>
    </div>
</body>
</html>
"""

# 生成圖表并轉(zhuǎn)換為base64
import base64

# 銷售趨勢圖
plt.figure(figsize=(12, 6))
sales_by_date.plot()
plt.title('每日銷售趨勢')
plt.xlabel('日期')
plt.ylabel('銷售額')
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
sales_trend_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 產(chǎn)品類別銷售圖
plt.figure(figsize=(12, 6))
sales_by_category.plot(kind='bar')
plt.title('各產(chǎn)品類別銷售額')
plt.xlabel('產(chǎn)品類別')
plt.ylabel('銷售額')
plt.xticks(rotation=45)
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
category_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 產(chǎn)品銷售圖
plt.figure(figsize=(12, 6))
top_10_products.plot(kind='bar')
plt.title('銷售額前10的產(chǎn)品')
plt.xlabel('產(chǎn)品名稱')
plt.ylabel('銷售額')
plt.xticks(rotation=45)
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
product_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 客戶消費分布圖
plt.figure(figsize=(12, 6))
sns.histplot(customer_purchase['總消費'], bins=20)
plt.title('客戶總消費分布')
plt.xlabel('總消費')
plt.ylabel('客戶數(shù)')
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
customer_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 支付方式分布圖
plt.figure(figsize=(10, 6))
plt.pie(payment_method_count, labels=payment_method_count.index, autopct='%1.1f%%')
plt.title('支付方式分布')
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
payment_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 銷售預(yù)測圖
plt.figure(figsize=(12, 6))
plt.plot(sales_series.index, sales_series.values, label='實際銷售額')
plt.plot(forecast.index, forecast.values, label='預(yù)測銷售額', linestyle='--')
plt.title('銷售額預(yù)測')
plt.xlabel('日期')
plt.ylabel('銷售額')
plt.legend()
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
forecast_chart = base64.b64encode(buf.read()).decode('utf-8')
plt.close()

# 產(chǎn)品關(guān)聯(lián)表
product_pairs_table = ''
for _, row in product_pairs_df.head(10).iterrows():
    product_pairs_table += f"<tr><td>{row['product1']}</td><td>{row['product2']}</td><td>{row['count']}</td></tr>"

# 填充報告數(shù)據(jù)
total_orders = len(df)
start_date = df['order_date'].min().strftime('%Y-%m-%d')
end_date = df['order_date'].max().strftime('%Y-%m-%d')
total_sales = df['total_amount'].sum()
total_customers = df['customer_id'].nunique()
total_products = df['product_name'].nunique()
top_category = sales_by_category.index[0]
top_category_percent = (sales_by_category.iloc[0] / total_sales) * 100
top_product = top_10_products.index[0]
top_product_sales = top_10_products.iloc[0]
top_customer_percent = (top_customers.sum() / customer_clv.sum()) * 100
top_payment_method = payment_method_count.idxmax()
top_payment_percent = (payment_method_count.max() / payment_method_count.sum()) * 100

# 生成最終報告
final_report = report_html.format(
    total_orders=total_orders,
    start_date=start_date,
    end_date=end_date,
    total_sales=total_sales,
    total_customers=total_customers,
    total_products=total_products,
    top_category=top_category,
    top_category_percent=top_category_percent,
    top_product=top_product,
    top_product_sales=top_product_sales,
    top_customer_percent=top_customer_percent,
    top_payment_method=top_payment_method,
    top_payment_percent=top_payment_percent,
    sales_trend_chart=sales_trend_chart,
    category_chart=category_chart,
    product_chart=product_chart,
    customer_chart=customer_chart,
    payment_chart=payment_chart,
    product_pairs_table=product_pairs_table,
    forecast_chart=forecast_chart
)

# 保存報告
with open('sales_analysis_report.html', 'w', encoding='utf-8') as f:
    f.write(final_report)

print("分析報告已生成：sales_analysis_report.html")