top of page

Data Scientist Program


Free Online Data Science Training for Complete Beginners.

No prior coding knowledge required!

Pandas' Techniques for Data Manipulation

# importing pandas as pd
import pandas as pd
import numpy as np
from pandas.api.types import CategoricalDtype

# creating a dataframe
df = pd.DataFrame({'A': ['John', 'Boby', 'Mina', 'Peter', 'Nicky'],
  'B': ['Masters', 'Graduate', 'Graduate', 'Masters', 'Graduate'],
  'C': [27, 23, 21, 23, 24]})


While pivot() provides general-purpose pivoting with various data types (strings, numerics, etc.), pandas also provide pivot_table() for pivoting with aggregation of numeric data. The function pivot_table() can be used to create spreadsheet-style pivot tables.

It takes a number of arguments:

  • Data: a DataFrame object.

  • values: a column or a list of columns to aggregate.

  • columns: a column, Grouper, an array that has the same length as data, or a list of them. Keys to group by on the pivot table column. If an array is passed, it is being used as the same manner as column values.

  • aggfunc: function to use for aggregation, defaulting to numpy. mean.

# Creates a pivot table dataframe
table = pd.pivot_table(df, values ='A', index =['B', 'C'],
                         columns =['B'], aggfunc = np.sum)

Iterating through pandas objects is generally slow. In many cases, iterating manually over the rows is not needed and can be avoided with one of the following approaches(in order as per performance ).

1. Vectorization :

many operations can be performed using built-in methods or NumPy functions or (boolean) indexing.

2. List Comprehensions:

3. df.apply():

4. Iter__family [df.itertuples / df.iteritems / df.itterrows]

Performance for iteration operations

#iterate using victorization
df_grad_vec = df['B'] == 'Graduate'

#iterate using list comprehensions
df_grad_comp = [x == 'Graduate' for x in df['B']]

#iterate using apply()
df_grad_apply = df.apply(lambda x:df['B']== 'Graduate' )

3. df.groupby (split / apply / combine)

    df.groupby(by=None, axis=0, as_index=True, dropna=True)

It takes a number of arguments:

  • by : mapping, function, label, or list of labels

  • axis {0 or ‘index’, 1 or ‘columns’}, default 0

  • as_index {bool, default True} :for aggregated output, return object with group labels as the index.

  • dropna {bool, default True} : if group keys contain NA values, NA values together with row/column will be dropped

By “group by” we are referring to a process involving one or more of the following steps:

1. Splitting the data into groups based on some criteria.

2. Applying a function to each group independently as the following:

2.1. Aggregation: compute a summary statistic (or statistics) for each group.

2.2. Transformation: perform some group-specific computations and return a like-indexed object.

2.3. Filtration: discard some groups, according to a group-wise computation that evaluates True or False.

3. Combining the results into a data structure.

df_grad_groupby = df.groupby(by='B', as_index=True)['C'].mean()

4. Categorical data

Categorical are a pandas data type corresponding to categorical variables in statistics. A categorical variable takes

on a limited, and usually fixed, number of possible values. Examples are gender, social class, blood type.

4.1. Controlling behaviors

the default behaviors is:

  • Categories are inferred from the data

  • Categories are unordered.

To control these behaviors we could use either

  • df.astype(‘category’)

  • passing argument (dtype=’category’) while creating the DataFrame

  • using CategoricalDtype.

4.2. Statistics for Categorical data(frequencies/proportions):
  • statistic for categorical data using describe()

  • statistic for categorical data using value_counts()

5.1. Covariance

to compute pairwise covariances among the series in the DataFrame


5.2. Correlation

to compute pairwise correlation among the series in the DataFrame


df3= pd.DataFrame(np.random.randn(1000, 5), columns=["a", "b", "c", "d", "e"])


Recent Posts

See All