Python 学习笔记

Python 学习笔记

pandas library

Day #1

df = pd.read_csv(’ ')

df.head(n)

df.tail(n)

df.nlargest(n,‘name’)

df.nsmallest(n,‘name’)

df [Pd[‘bla bla’]==‘labalaba’]

Day #2

df.shape

df.columns.tolist( )

df[‘name’].unique( )

df = df.replace(‘Orig’ , ‘New’)

df.isnull( ) // Checks if each value in the dato frame is missing

Trick: print(df.isnull().sum()) // return the total amount of missing values for each feature(column)

NaN stands for ‘not a number’

creating columns:

df[‘new column name’] = fomula

e.g. df[’% Female’] = df[‘Female’] / df[‘Total’]

df.dtypes

df.select_dtypes([‘object’])

measures of spread or distribution

df[‘column name’].mean( )

df[‘column name’].median( )

df[‘column name’].mode( )

df.groupby(by=‘column name’).agg(‘mean’)

// agg stands for ‘aggregate’. group the data by the ‘column name’

df[‘column name’].max( ) — df[‘column name’].min( )

df[‘column name’].std( ) // standard deviation

df[‘column name’].var( ) // variance

df.groupby(by=‘column name’).agg(‘std’)

correlation coefficient

df.corr( )

df.corr( )[‘x’] [‘y’]

Day #3

Bar Chart

df_top10 = df.nlargest(10,‘Total’)

df_top10.plot.bar(x=‘Major Name’, y=‘Total’)

Histograms

df.hist(column = ‘% Female’)

Boxplot

df.boxplot(column = ‘% Female’, vert=False)

Scatterplots

df.plot.scatter(x=‘Male’, y=‘Female’)

Simulation

import random
random.ranint(1,10)

for i in range(1,100):
  print(random.randint(1,10))

Tricks

exercise 1

通过添加一个新的布尔值列（当且仅当以下两项均为 True 时为 True）修改 cities 表格：

• 城市以圣人命名。
• 城市面积大于 50 平方英里。

cities['wide and saint'] = (cities['Area square miles'] > 50) & cities['City name'].apply(lambda name: name.startswith('San'))
cities