Pandas operations

pandas.md

Importing Data

• Dictionary to DataFrame

df = pd.Dataframe(my_dict)

• Dictionary to DataFrame with row labels

row_labels = ['label0', 'label1']
df.index = row_labels

• csv to DataFrame

from csv file  
`df = pd.read_csv(file_name)

• csv to DataFrame with row labels

df = pd.read_csv(file_name, index_col=0)

Indexing and Selecting Data

• Get column(s) as Series

s = df['col_1']

• Get column(s) as DataFrame

d1 = df[['col_1']]
d2 = df[['col_1', 'col2']]

• Access by slicing

d = df[1:4]

• Access by label: loc

• Access by position: iloc

• Access by boolean arrays

Filter, Modify, Impute DataFrame

• Filtering DataFrame

• Add new column to DataFrame by using another column:

df[lenght_column_a] = df['column_a'].apply(len)
df[capitalized_column_b] = df['column_b'].apply(str.upper)

• TODO

pd.to_numeric(df['Column Name'], errors='coerce')

• Change data type of DataFrame column

df.column_name = df.column_name.astype(int)

• TODO

df.drop(['columnheading1', 'columnheading2'], axis=1, inplace=True)

• Impute Missing Values with 0

df.update(df[['a','b','c']].fillna(0))

Configurations

pandas.set_option('display.width', 100)
pandas.set_option('precision', 3)

Statistics

df.shape 
df.dtypes
df.info()
df.describe
df.corr(method='pearson')
df.skew() #Values closer to zero show less skew.