Data Wrangling is the process of converting data from the initial format to a format that may be readable and better for analysis.

Here we use the below data set :

https://archive.ics.uci.edu/ml/machine-learning-databases/autos/imports-85.data

Import pandas

Open Jupyter notebook or any online jupyter notebook editor and import pandas-

import pandas as pd
import matplotlib.pylab as plt

Want to add a caption to this image? Click the Settings icon.

Reading the data and add header

filename = "https://s3-api.us-geo.objectstorage.softlayer.net/cf-courses-data/CognitiveClass/DA0101EN/auto.csv"

headers = ["symboling","normalized-losses","make","fuel-type","aspiration", "num-of-doors","body-style", "drive-wheels","engine-location","wheel-base",

"length","width","height","curb-weight","engine-type", "num-of-cylinders", "engine-size","fuel

-system","bore","stroke","compression-ratio","horsepower", "peak-rpm","city-mpg","highway-mpg","price"]

Want to add a caption to this image? Click the Settings icon.

Read CSV

df = pd.read_csv(filename, names = headers) 

Show data in tabular form

df.head()

Data display in tabular form and you will face some challenges like this-

• identify missing data

• deal with missing data

• correct data format

Identify and handle missing values

Identify missing values

Convert "?" to NaN

Missing data comes with the question mark "?". We replace "?" with NaN (Not a Number)

Example:

import numpy as np
# replace "?" to NaN
df.replace("?", np.nan, inplace = True)
df.head(5)

It set NaN at first five index row where "?" is presented.

How to detect missing data:

There are two method used to detect missing data.

• .isnull() - Return true at the place of missing data and other place return false.

• .notnull() - Return true at the placed data and false at missing data place.

Example:

mis_value = df.isnull()
mis_value.head(5)

Count missing value -In columns

Using for loop:

Example:

Write this for loop and find result

for column in mis_value .columns.values.tolist():
print(column)
print (mis_value [column].value_counts())
print("")

How we will work with missing data

Drop data

• drop the whole row- Let suppose any value is necessary like price but it is missing at any row then we remove whole row.

• drop the whole column - let we suppose if price is missing at any column then it reason of delete whole column because price is necessary for data science to calculate price.

Replace data

• replace it by mean

• replace it by frequency - replace as per frequency for example- 84 % is good, and 16% bad, then 16% remove by good.

• replace it based on other functions

Calculate the average of any column

Example

• avg= df["column name"].astype("float").mean(axis=0)

• print("Average of column name:", avg)

Replace "NaN" by mean value - of any column

Example

df["column_name"].replace(np.nan, avg, inplace=True)

Calculate the mean value - of any column

Example

avg=df['column_name'].astype('float').mean(axis=0)
print("Average of column_name:", avg)

Replace NaN by mean value

Example

df["column_name"].replace(np.nan, avg, inplace=True)

How count each column data separately

Use value_counts() function

Example:

df['column_name'].value_counts()

Output like this: let suppose column_name is qualification then count each qualification with name.

mca 78

bca 45

Calculate for us the most common (max) automatically

df['column_name'].value_counts().idxmax()

Output:

mca 78

Replace NaN by most frequent

Example

df["column_name"].replace(np.nan, "four", inplace=True)

All NaN replace by most frequent- by "four"

Drop whole row with NaN in "Column_name" column

Let suppose column_name is "price"

df.dropna(subset=["price"], axis=0, inplace=True)

# reset index, because we dropped two rows

df.reset_index(drop=True, inplace=True)

Correct data format

In Pandas, we use

• .dtype() to check the data type

• .astype() to change the data type

Show list of data type:

Use this syntax to list data type -

df.dtypes

How to convert data type in proper format

There are different type of data format used -

Syntax

df[["column1", "column2"]] = df[["column1", "column2"]].astype("float")
df[["column3"]] = df[["column3"]].astype("int")
df[["column4"]] = df[["column4"]].astype("float")
df[["column5"]] = df[["column5"]].astype("float")

Again check it by using following -

It show list so that you can verify that data type is change or not

Syntax:

df.dtypes

Data Standardization

What is Standardization?

Standardization is the process of transforming data into a common format which allows the researcher to make the meaningful comparison.

Example

Transform mpg to L/100km

The formula for unit conversion is

L/100km = 235 / mpg

First go through the data to verify it by using this syntax-

Syntax:

df.head()

Example:

Convert mpg to L/100km by mathematical operation

df['city-L/100km'] = 235/df["city-mpg"]

It add new column city-L/100km after change the value of column city-mpg

# check your transformed data

df.head()

Data Normalization

Why normalization?

Normalization is the process of transforming values of several variables into a similar range.

Example:

# replace (original value) by (original value)/(maximum value 
df['length'] = df['length']/df['length'].max()
df['width'] = df['width']/df['width'].max()

Binning

Why binning?

Binning is a process of transforming continuous numerical variables into discrete categorical 'bins', for grouped analysis.

Indicator variable (or dummy variable)

What is an indicator variable?

An indicator variable (or dummy variable) is a numerical variable used to label categories. They are called 'dummies' because the numbers themselves don't have inherent meaning.

Why we use indicator variables?

So we can use categorical variables for regression analysis in the later modules.