DataFrame() constructor

Dataframe is a 2-D Array.
Used to create or reconstruct a DataFrame object

DataFrame(data=None, index=None, columns=None, dtype=None, copy=None)
      

any()

check if at least one value in a series or along an axis evaluates to True.

# data.csv
Duration,Pulse,Maxpulse,Calories
45,       117,    148,  406.0
60,       102,    127,  300.5
60,       110,    136,
45,       104,     134, 253.3

import pandas as pd
df = pd.read_csv('data.csv')
all_cols = [col for col in df.columns] 
print(all_cols)       # [Duration,Pulse,Maxpulse,Calories]

cols_with_missing = [col for col in df.columns if df[col].isnull().any()]
print(cols_with_missing)  # [Calories]

describe()

describe(): Returns 8 values for each coloumn
  count: The number of non-null values. (For Duration coloumn, count=3)
  mean: The average value. (For Duration coloumn, 70+60+50/3 = 60)
  std: The standard deviation.
  min: The minimum value. (For duration coloumn, min val=50)
  25%: The 25th percentile (first quartile).
   Imagine sorting each column from lowest to highest value. If you go a quarter way through the list, you'll find a number that is bigger than 25% of the values, ie 25%
  50%: The 50th percentile (median).
  75%: The 75th percentile (third quartile).
  max: The maximum value. (For duration coloumn, max val=70)

# data.csv 
Duration,   Pulse,  Maxpulse,   Calories
60,         110,    130,        409.1
60,         117,    145,        479.0
60,         103,    135,        340.0
,         109,    175,        282.4

import pandas as pd
df = pd.read_csv('data.csv')
print(df.describe())

       Duration       Pulse    Maxpulse    Calories
count       3.0    4.000000    4.000000    4.000000
mean       60.0  109.750000  146.250000  377.625000
std        10.0    5.737305   20.155644   85.148904
min        50.0  103.000000  130.000000  282.400000
25%        55.0  107.500000  133.750000  325.600000
50%        60.0  109.500000  140.000000  374.550000
75%        65.0  111.750000  152.500000  426.575000
max        70.0  117.000000  175.000000  479.000000
      

drop() and dropna()

dropna(): Cleans out missing data (NaN/None). DataFrame with NA entries dropped from it or None if inplace=True.
drop(): Deletes explicit rows or columns by name.

# Drops the column named 'Age'
df_dropped = df.drop(columns=['Age']) 

# Drops the row with index label 2
df_dropped = df.drop(index=2) 

Example:
data_frame = 
       Duration       Pulse    Maxpulse    Calories
count       3.0    4.000000    4.000000    4.000000
mean       60.0  109.750000  146.250000  377.625000
std        10.0    5.737305   20.155644   85.148904
min        50.0  103.000000  130.000000  282.400000
25%        55.0  107.500000  133.750000  325.600000
50%        60.0  109.500000  140.000000  374.550000
75%        65.0  111.750000  152.500000  426.575000
max        70.0  117.000000  175.000000  479.000000

// Pulse col will be removed
d = data_frame.drop(['Pulse'], axis=1)
          

data.csv
Duration,Pulse,Maxpulse,Calories
45,       117,  148,    406.0
60,       102,  127,    300.5
60,       110,  136,
45,       104,  134,    253.3
                  

import pandas as pd
df = pd.read_csv('data.csv')
print(df.dropna())  # OR df.dropna(axis=0)

   Duration  Pulse  Maxpulse  Calories
0        45    117       148     406.0
1        60    102       127     300.5
3        45    104       134     253.3
                  

print(df.dropna(axis=0,subset=['Pulse']))
Duration,Pulse,Maxpulse,Calories
45,       117,  148,    406.0
60,       102,  127,    300.5
60,       110,  136,    NaN
45,       104,  134,    253.3
                  

head()

head returns 1st n rows from dataframe, default is 5

isNull

Creates a dataframe with 1 or 0. if value is present, take it as 0, if NaN or None take it as 1

X_train
rollno  marks
5       NaN
NaN     NaN
10      20

t = (X_train.isnull())
rollno = [0, 1, 0]
marks =  [1, 1, 0]
      

read_csv(), .columns

read_csv(): read the CSV file and create a Dataframe from it
.columns: list the coloumn header in the Dataframe

# data.csv 
Duration,   Pulse,  Maxpulse,   Calories
60,         110,    130,        409.1
60,         117,    145,        479.0
60,         103,    135,        340.0
,         109,    175,        282.4

$ test.py
import pandas as pd
df = pd.read_csv('data.csv')    #read in data frame
print(df.to_string())
print(df.columns)

   Duration  Pulse  Maxpulse  Calories
0      60.0    110       130     409.1
1      60.0    117       145     479.0
2      60.0    103       135     340.0
3       NaN    109       175     282.4

Index(['Duration', 'Pulse', 'Maxpulse', 'Calories'], dtype='object')
      

Features

Selecting some columns from dataframe is called features.

import pandas as pd
df = pd.read_csv('data.csv')
print(df.to_string())
features = ['Duration', 'Pulse']
print(df[features])

    Duration  Pulse  Maxpulse  Calories
0      60.0    110       130     409.1
1      60.0    117       145     479.0
2      60.0    103       135     340.0
3       NaN    109       175     282.4

   Duration  Pulse
0      60.0    110
1      60.0    117
2      60.0    103
      

select_dtypes()

// select_dtypes() filter columns based on their data type.
DataFrame.select_dtypes(include=None, exclude=None)

//exclude=['object']: Give me every column except those labeled as object.
X = melb_predictors.select_dtypes(exclude=['object'])

What is dtype

shape()

Find row, coloumn count of datafram read

import pandas as pd
X_full = pd.read_csv('train.csv', index_col='Id')
rows, cols = X_full.shape
print(f"Rows: {rows}, Columns: {cols}")   //Rows: 1460, Columns: 80
      

sum()

Calculate total sum

Feature A: [False, True, False]
Feature B: [True, True, False]

.sum() calculates:
Feature A: 0 + 1 + 0 = 1
Feature B: 1 + 1 + 0 = 2