Pandas数据分析

Pandas数据分析

1.基本统计函数

函数	说明
sum()	求和
count()	统计个数
max()	求最大值
min()	求最小值
median()	求中位数
mean()	求平均值
mode()	求众数
var()	求方差
std()	求标准差
quantile()	求分位数

(1)sum()

df.sum(axis=0或1,numeric_only=布尔值)

numeric_only的默认值为False，表示对所以的行或列进行求和

True则会对数字进行求和

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.sum(axis=0, numeric_only=False)
print(res)

(2)count()

df.count(axis=0或1)

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.count(axis=0)
print(res)

(3)max() & min()

df.max(axis=0或1,numeric_only=布尔值)
df.min(axis=0或1,numeric_only=布尔值)

numeric_only的默认值为False，表示对所以的行或列进行求最值

True则会对数字进行求最值

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.max(axis=0, numeric_only=True)
print(res)

(4)median()

df.median(axis=0或1,numeric_only=布尔值)

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.median(axis=0, numeric_only=True)
print(res)

(5)mode()

df.mode(axis=0或1, numeric_only=布尔值)

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.mode(axis=0, numeric_only=False)
print(res)

(6)quantile()

df.quantile(axis=0或1,q=值)

q是一个浮点数表示取多少百分位数

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.quantile(axis=0, numeric_only=True, q=0.5)
print(res)

2.其他统计函数

函数	说明
unique()	统计取值的种类
value_counts()	统计取值个数
pct_change()	求变化百分比
idxmax()	求最大值的行名
idxmin()	求最小值的行名

(1)unique()

df[列名].unique()

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df["平均价格（元/斤）"].unique()
print(res)

返回的结果是个列表，里面是这一列全部的取值

(2)value_counts()

df[列名].value_counts()

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df["平均价格（元/斤）"].value_counts()
print(res)

返回结果为Series，Series的index是值的种类，value是值对应的个数

(3)pct_change()

df.pct_change(axis=0或1)

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data).iloc[:, [1, 2, 3]]
res = df.pct_change(axis=0)
print(res)

这个函数将每一个元素和前面的值进行比较，计算变化百分比

(4)idxmax() & idxmin()

df[列名].idxmax()
df[列名].idxmin()

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df["月销量（斤）"].idxmax()
print(res)

3.整体情况

(1)describe()

describe()函数会一次性获得一个数据集

df.describe()

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.describe()
print(res)

(2)info

info()函数会的一个数据集，主要为类型，列名等

df.info()

import pandas as pd

data = {
    "水果名称": ["苹果", "香蕉", "橙子", "西瓜", "草莓", "葡萄", "芒果", "菠萝", "梨", "桃子"],
    "平均价格（元/斤）": [5, 3, 6, 2, 10, 8, 7, 4, 4, 5],
    "月销量（斤）": [500, 350, 400, 600, 200, 300, 250, 180, 220, 150],
    "维生素 C 含量（mg/100g）": [4, 8, 30, 6, 60, 5, 28, 18, 4, 7]
}
df = pd.DataFrame(data)
res = df.info()
print(res)

4.聚合函数

聚合函数可以实现多个对象进行统计

df.agg(列表)

5.数据分类

(1)创建分组

df.groupby(列名或列表)

import pandas as pd

data = {
    "学生姓名": ["张三", "李四", "王五", "赵六", "孙七"],
    "班级": ["一班", "二班", "三班", "一班", "二班"],
    "性别": ["男", "女", "男", "女", "男"],
    "年龄": [18, 19, 20, 17, 18]
}

df = pd.DataFrame(data)
group = df.groupby("班级")
for i in group:
    print(i)

(2)统计分析

可以使用一些统计函数进行分析

import pandas as pd

data = {
    "学生姓名": ["张三", "李四", "王五", "赵六", "孙七"],
    "班级": ["一班", "二班", "三班", "一班", "二班"],
    "性别": ["男", "女", "男", "女", "男"],
    "年龄": [18, 19, 20, 17, 18]
}

df = pd.DataFrame(data)
group = df.groupby("班级")
res = group.count()
print(res)