Series and Dataframe

Series

• A Series is a one-dimensinal object

• It can hold any data type

• Series consists of index label and value

• Value can be accessed via index label

• Index can be duplicated

Simple series

import pandas as pd

x = pd.Series([1,2,3])
x

index   values
--------------
0       1
1       2
2       3

x[0]
1

Define new index

x = pd.Series([1,2,3], index=['a','b','c'])
x

index   values
--------------
a       1
b       2
c       3

x['a']
1

Dictionary type series

data = {'abc': 1, 'def': 2, 'xyz': 3}
x = pd.Series(data)
x

index   values
--------------
abc     1
def     2
xyz     3

x['abc']
1

Scalar value series

The value gets repeated for each of the indexed defined.

x = pd.Series(1, index=['a','b','c'])
x

index   values
--------------
a       1
b       1
c       1

Dataframe

• A Dataframe is a two dimensinal object that can have columns

• Dictionaries, lists, series can be included

• Most commonly used pandas object

Dataframe with Numpy

import numpy as np
import pandas as pd

dates = pd.date_range('20200226', periods=3)
# DatetimeIndex(['2020-02-26', '2020-02-27', '2020-02-28'], dtype='datetime64[ns]', freq='D')
columns = list('ABC')
['A', 'B', 'C']

data = np.random.randn(3,3)
# array([[-0.10914734, -0.75659384, -0.06899813],
#        [ 1.37714538,  2.09279708, -0.05586049],
#        [ 0.2282605 , -1.54231927, -0.34941844]])

df = pd.DataFrame(data, index=dates, columns=columns)
df


index       A            B            C
--------------------------------------------
2020-02-26    0.109953    -0.134801    0.023890
2020-02-27    1.417591    0.800834    0.145955
2020-02-28    -1.428734    0.438276    0.422585

Get Summary of our data

df.describe()

         A             B            C
-----------------------------------------
count     3.000000     3.000000    3.000000
 mean    -0.297529    -0.695912    0.751660
  std     0.966246     1.045112    0.839380
  min    -1.155043    -1.624613    -0.167104
  25%    -0.820992    -1.261776    0.388307
  50%    -0.486941    -0.898940    0.943717
  75%     0.131229    -0.231561    1.211043
  max     0.749398     0.435818    1.478368

Get cumulative sum

df.apply(np.cumsum)


            A             B             C 
---------------------------------------------
2020-02-26    -0.076596    -0.470757     0.109522
2020-02-27     1.390033    -0.875267    -1.120344
2020-02-28     1.504389    -0.522906    -3.327057

Referenced site

• Series and DataFrame in Python