Pandas - DataFrame¶

DataFrames description:

two dimensional labeld data
supports diffrent data types
easy to manipulate: eg. reshaping, slicing, grouping

used for:

data wrangling
visualization
data processing
exploratory data analysis
creating models

Create DataFrame by dict¶

In [2]:

import pandas as pd
data = {'age':  [22,55,43],
        'names': ['A','B','C'],
        'country': ['uk', 'us','de'],
        }

df = pd.DataFrame (data)
print(df)

   age names country
0   22     A      uk
1   55     B      us
2   43     C      de

Stats about DataFrame¶

In [3]:

print(df.describe())
print(df.info())
print(df.shape)

             age
count   3.000000
mean   40.000000
std    16.703293
min    22.000000
25%    32.500000
50%    43.000000
75%    49.000000
max    55.000000
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3 entries, 0 to 2
Data columns (total 3 columns):
age        3 non-null int64
names      3 non-null object
country    3 non-null object
dtypes: int64(1), object(2)
memory usage: 152.0+ bytes
None
(3, 3)

Use specific columns¶

In [4]:

df = pd.DataFrame (data, columns = ['age','country'])
print(df)

   age country
0   22      uk
1   55      us
2   43      de

Filter Rows¶

In [5]:

dfFiltered = df[df['age']<50]
print(dfFiltered)

   age country
0   22      uk
2   43      de

numpy array <> pandas DataFrame¶

You can use:

np.array(yourDataFrameVariable)
yourDataFrameVariable.values

to transform yout pandas DataFrame to a numpy array

In [6]:

import numpy as np
print(type(dfFiltered))
print(dfFiltered)

npArray = np.array(dfFiltered)
print(type(npArray))
print(npArray)

dfAgain = pd.DataFrame(npArray)
print(type(dfAgain))
print(dfAgain)

<class 'pandas.core.frame.DataFrame'>
   age country
0   22      uk
2   43      de
<class 'numpy.ndarray'>
[[22 'uk']
 [43 'de']]
<class 'pandas.core.frame.DataFrame'>
    0   1
0  22  uk
1  43  de

In [7]:

npArray = dfFiltered.values
print(type(npArray))
print(npArray)

<class 'numpy.ndarray'>
[[22 'uk']
 [43 'de']]