pandas練習（四）--- 應用Apply函數

探索學生對酒的消費狀況

數據見github

步驟1 - 導入必要的庫

import pandas as pd import numpy as np

步驟2 - 數據集

path4 = "./data/student-mat.csv"   

步驟3 將數據命名爲student

student = pd.read_csv(path4)
student.head()

輸出：

步驟4 從'school'到'guardian'將數據切片

stud_alcoh = student.loc[: , "school":"guardian"] stud_alcoh.head()

輸出：

 

步驟5 建立一個捕獲字符串的lambda函數

captalizer = lambda x: x.upper()

步驟6 使'Fjob'列都大寫

stud_alcoh['Fjob'].apply(captalizer)

輸出：

0       TEACHER
1         OTHER
2         OTHER
3      SERVICES
4         OTHER
5         OTHER
6         OTHER
7       TEACHER
8         OTHER
9         OTHER
10       HEALTH
11        OTHER
12     SERVICES
13        OTHER
14        OTHER
15        OTHER
16     SERVICES
17        OTHER
18     SERVICES
19        OTHER
20        OTHER
21       HEALTH
22        OTHER
23        OTHER
24       HEALTH
25     SERVICES
26        OTHER
27     SERVICES
28        OTHER
29      TEACHER
         ...   
365       OTHER
366    SERVICES
367    SERVICES
368    SERVICES
369     TEACHER
370    SERVICES
371    SERVICES
372     AT_HOME
373       OTHER
374       OTHER
375       OTHER
376       OTHER
377    SERVICES
378       OTHER
379       OTHER
380     TEACHER
381       OTHER
382    SERVICES
383    SERVICES
384       OTHER
385       OTHER
386     AT_HOME
387       OTHER
388    SERVICES
389       OTHER
390    SERVICES
391    SERVICES
392       OTHER
393       OTHER
394     AT_HOME
Name: Fjob, dtype: object

步驟7 打印數據集的最後幾行元素

stud_alcoh.tail()

輸出：

步驟8 注意到原始數據框仍然是小寫字母，接下來改進一下

stud_alcoh['Mjob'] = stud_alcoh['Mjob'].apply(captalizer) stud_alcoh['Fjob'] = stud_alcoh['Fjob'].apply(captalizer) stud_alcoh.tail()

輸出：

步驟9 建立一個名爲majority的函數，它返回一個布爾值到一個名爲legal_drinker的新列（多數年齡大於17歲）

def majority(x): if x > 17: return True else: return False

stud_alcoh['legal_drinker'] = stud_alcoh['age'].apply(majority) stud_alcoh.head()

輸出：

步驟10 將數據集的每一個數字乘以10

def times10(x): if type(x) is int: return 10 * x return x

stud_alcoh.applymap(times10).head(10)

輸出：

參考連接：

一、http://pandas.pydata.org/pandas-docs/stable/cookbook.html#cookbook

二、https://www.analyticsvidhya.com/blog/2016/01/12-pandas-techniques-python-data-manipulation/

三、https://github.com/guipsamora/pandas_exercises