import pandas as pd
# 從CSV文件讀取用戶點(diǎn)擊數(shù)據(jù)
click_data = pd.read_csv('user_clicks.csv', parse_dates=['click_time'])
# 從MySQL數(shù)據(jù)庫(kù)讀取訂單數(shù)據(jù)
import pymysql
conn = pymysql.connect(host='localhost', user='root', password='123456', db='ecommerce')
order_data = pd.read_sql('SELECT * FROM orders WHERE order_date > "2025-01-01"', conn)

import requests
from bs4 import BeautifulSoup
headers = {'User-Agent': 'Mozilla/5.0'}
url = 'https://www.example.com/products'
response = requests.get(url, headers=headers)
soup = BeautifulSoup(response.text, 'html.parser')
products = []
for item in soup.select('.product-item'):
    products.append({
        'name': item.select_one('.name').text.strip(),
        'price': float(item.select_one('.price').text[1:])
    })
product_df = pd.DataFrame(products)

import requests
import json
api_url = 'https://www.zdaye.com'
response = requests.get(api_url)
sales_data = json.loads(response.text)['data']  # 轉(zhuǎn)換為Python字典
sales_df = pd.DataFrame(sales_data)

# 檢查缺失值
print(user_data.isnull().sum())
# 年齡缺失用中位數(shù)填充（抗異常值）
user_data['age'].fillna(user_data['age'].median(), inplace=True)
# 地址缺失用眾數(shù)填充（常見(jiàn)值）
most_common_city = user_data['city'].mode()[0]
user_data['city'].fillna(most_common_city, inplace=True)

Q1 = order_data['amount'].quantile(0.25)
Q3 = order_data['amount'].quantile(0.75)
IQR = Q3 - Q1
lower_bound = Q1 - 1.5*IQR
upper_bound = Q3 + 1.5*IQR
abnormal_orders = order_data[(order_data['amount'] < lower_bound) | (order_data['amount'] > upper_bound)]

# 日期標(biāo)準(zhǔn)化
order_data['order_date'] = pd.to_datetime(order_data['order_date'], format='%Y-%m-%d')
# 金額單位轉(zhuǎn)換（元→千元）
order_data['amount_k'] = order_data['amount'] / 1000

print(sales_data.describe())
"""
              amount
count  12584.000000
mean     156.320000
std       48.750000
min       12.000000
25%      125.000000
50%      150.000000
75%      175.000000
max      320.000000
"""

import seaborn as sns
import matplotlib.pyplot as plt
corr_matrix = sales_data[['price', 'discount', 'amount']].corr()
sns.heatmap(corr_matrix, annot=True, cmap='coolwarm')
plt.title('銷售因素相關(guān)性分析')
plt.show()

from statsmodels.tsa.seasonal import seasonal_decompose
result = seasonal_decompose(sales_data.set_index('date')['amount'], model='additive')
result.plot()
plt.show()

# 計(jì)算RFM指標(biāo)
from datetime import datetime
today = datetime(2025,7,16).date()
rfm = order_data.groupby('user_id').agg({
    'order_date': lambda x: (today - x.max()).days,  # Recency
    'user_id': 'count',                             # Frequency
    'amount': 'sum'                                  # Monetary
})
rfm.columns = ['R', 'F', 'M']
# 標(biāo)準(zhǔn)化處理
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
rfm_scaled = pd.DataFrame(scaler.fit_transform(rfm), columns=rfm.columns)
# K-means聚類
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters=4, random_state=42)
rfm['cluster'] = kmeans.fit_predict(rfm_scaled)

from statsmodels.tsa.arima.model import ARIMA
# 訓(xùn)練集/測(cè)試集劃分
train = sales_data[:100]
test = sales_data[100:]
# 模型擬合
model = ARIMA(train['amount'], order=(1,1,1))
model_fit = model.fit()
# 預(yù)測(cè)
forecast = model_fit.forecast(steps=len(test))
# 評(píng)估
from sklearn.metrics import mean_squared_error
mse = mean_squared_error(test['amount'], forecast)
print(f'預(yù)測(cè)均方誤差: {mse:.2f}')

from sklearn.neighbors import NearestNeighbors
# 構(gòu)建用戶-商品矩陣
user_item_matrix = pd.pivot_table(click_data, values='click', index='user_id', columns='product_id', aggfunc='count').fillna(0)
# 訓(xùn)練模型
model = NearestNeighbors(n_neighbors=5, metric='cosine')
model.fit(user_item_matrix)
# 為用戶推薦商品
user_id = 123
distances, indices = model.kneighbors([user_item_matrix.loc[user_id]])
recommended_products = user_item_matrix.columns[indices[0][1:]]  # 排除自身

# 銷售額趨勢(shì)圖
plt.figure(figsize=(12,6))
sales_data.set_index('date')['amount'].plot()
plt.title('日銷售額趨勢(shì)')
plt.ylabel('金額（千元）')
plt.grid(True)
plt.show()
 
# 商品類別分布
plt.figure(figsize=(8,6))
sales_data['category'].value_counts().plot(kind='barh')
plt.title('商品類別銷售占比')
plt.xlabel('銷量')
plt.show()

# 熱力圖展示不同時(shí)段銷售情況
pivot_table = sales_data.pivot_table(index='hour', columns='weekday', values='amount', aggfunc='sum')
sns.heatmap(pivot_table, cmap='YlOrRd', annot=True, fmt='.0f')
plt.title('工作日/時(shí)段銷售熱力圖')
plt.show()
 
# 地理分布圖（需安裝folium）
import folium
m = folium.Map(location=[35,105], zoom_start=4)
for _, row in store_data.iterrows():
    folium.CircleMarker(
        location=[row['lat'], row['lng']],
        radius=row['sales']/1000,
        color='red',
        fill=True
    ).add_to(m)
m.save('stores_map.html')

import plotly.express as px
fig = px.scatter(sales_data, x='price', y='amount', 
                 color='category', size='quantity',
                 hover_data=['product_name'],
                 title='商品價(jià)格-銷量關(guān)系分析')
fig.show()

from apscheduler.schedulers.blocking import BlockingScheduler
def generate_daily_report():
    # 數(shù)據(jù)采集、分析、可視化代碼
    report = create_sales_report()
    report.to_excel('daily_report.xlsx')
scheduler = BlockingScheduler()
scheduler.add_job(generate_daily_report, 'cron', hour=8, minute=30)
scheduler.start()

from flask import Flask, request, jsonify
app = Flask(__name__)
@app.route('/analyze', methods=['POST'])
def analyze():
    data = request.json
    # 執(zhí)行分析邏輯
    result = {
        'trend': calculate_trend(data),
        'segments': cluster_users(data)
    }
    return jsonify(result)
if __name__ == '__main__':
    app.run(port=5000)

# lambda_function.py
import pandas as pd
def lambda_handler(event, context):
    # 從S3讀取數(shù)據(jù)
    s3_client = boto3.client('s3')
    obj = s3_client.get_object(Bucket='my-data-bucket', Key='sales.csv')
    data = pd.read_csv(obj['Body'])
    # 執(zhí)行分析
    result = analyze_data(data)
    # 存儲(chǔ)結(jié)果
    s3_client.put_object(Bucket='my-result-bucket', 
                        Key='analysis_result.json',
                        Body=result.to_json())
    return {'statusCode': 200}