Introduction to Pandas

Python Library : Pandas

PANDAS : Open source data analysis library written in python. It leverages the power and speed of numpy to make data analysis and EDA (Exploratory Data Analysis) really easy task for any data scientist.

Analogy : As we took vegetable from the market cleaning before cooking it, that's what pandas do (cleaning of the present data)

---

Feature means : Columns

Records means : Rows

---

Importing pandas and numpy libraries :

import pandas as pd
import numpy as np

Importing Dataset from local directory

data = pd.read_csv("/train.csv")

---

Dataset Datatype : DataFrame which is basically tabular form of data

type(data) #data structure type

Output :

---

---

How many null values are available in a given dataset

data.isnull().sum() # finds sum of all features(column) null values count

Output :

dtype: int64

---

Structure of the Dataset in given row and column

data.shape # gives row and column

Output :

(891, 12) #891 records, 12 features

---

Dictionary to DataFrame data type

Creating Dictionary :

dic1 = {"name": ["Priya", "Nikhil", "Sanjay"], "age": [25, 25, 37], "city": ["chandigarh", "Bengaluru", "Delhi"]}

Print dic1 dictionary:

dic1

Output :

{'name': ['Priya', 'Nikhil', 'Sanjay'],
 'age': [25, 25, 37],
 'city': ['chandigarh', 'Bengaluru', 'Delhi']}

Converting dictionary to DataFrame

dataFrame = pd.DataFrame(dic1) #convert dic1 to DataFrame

Printing dataFrame

dataFrame

Output :

---

List of all features present in a given dataset

data.columns

Output :

Index(['PassengerId', 'Survived', 'Pclass', 'Name', 'Sex', 'Age', 'SibSp',
       'Parch', 'Ticket', 'Fare', 'Cabin', 'Embarked'],
      dtype='object')

---

Every Column's datatypes

data.dtypes # for every features datatypes

Output :

dtype: object

---

Gives All Information about Dataset with respect to numerical features

data.info # information about dataset

Output :

dtype: float64

---

Series and DataFrame Data Structure

Data type of given code:

type(data['Age']) #gives datatype of 'age' , data type is Series because this is one dimentional

Output : pandas.core.series.Series

891 rows × 1 columns

---

Data type representation of given code:

type(data[["Name", "Age","Ticket"]] ) # data type is DataFrame , because this is greater than one dimensional data.

Output : pandas.core.frame.DataFrame

891 rows × 3 columns

---

Difference between loc and iloc:

Row indexing : Preview 0th and 4th records only with all features

data.loc[[0,4]]

Output :

---

---

Specific row and column indexing : repesent records and features with a given range, --excluding last record and feature

data.iloc[2:5, 0:4]

Output :