import pandas as pd
from sklearn.decomposition import PCA
from matplotlib import pyplot as plt
# 读取4张表的数据
aisles = pd.read_csv('C:\\Users\\RSB\\Desktop\\python\\aisles.csv')
prior = pd.read_csv('C:\\Users\\RSB\\Desktop\\python\\order_products__prior.csv')
orders = pd.read_csv('C:\\Users\\RSB\\Desktop\\python\\orders.csv')
products = pd.read_csv('C:\\Users\\RSB\\Desktop\\python\\products.csv')
#数据量太大，截取一部分作为研究
aisles1=aisles.loc[:50000,]
prior1=prior.loc[:50000,]
orders1=orders.loc[:50000,]
products1=products.loc[:50000,]
# 合并4张表到一张表 （用户-物品类别）
_mg = pd.merge(prior1, products1, on=['product_id', 'product_id'])
_mg = pd.merge(_mg, orders1, on=['order_id', 'order_id'])
mt = pd.merge(_mg, aisles1, on=['aisle_id', 'aisle_id'])
print(mt.head(10))
#交叉表，特殊的分组表
cross=pd.crosstab(mt['user_id'],mt['aisle'])
cross.head(10)
#主成份分析
pca=PCA(n_components=0.9)
data=pca.fit_transform(cross)

plt.scatter(data[:,0],data[:,1])
plt.show()