Data cleaning and preparation quiz

Question 1

What is "data cleaning" in machine learning?

Accepted Answer

Fixing errors, missing values, and duplicates so data is accurate

Answer

Deleting all the data and starting over

Answer

Making the computer run faster

Answer

Drawing colorful charts of the data

Question 2

Why do people say "garbage in, garbage out" when talking about data?

Accepted Answer

Bad input data leads to bad model results

Answer

Computers literally produce trash

Answer

You should throw away your computer

Answer

Garbage is good for training models

Question 3

What is a "missing value" in a dataset?

Accepted Answer

A cell where information should be but is empty

Answer

A value that is too large to store

Answer

A number written in the wrong color

Answer

A value that appears too many times

Question 4

What are two common ways to deal with missing values?

Accepted Answer

Remove the rows, or fill them with a sensible value

Answer

Ignore the whole dataset, or print it on paper

Answer

Always set them to zero, or always delete the file

Answer

Color them red, or make them bold

Question 5

What is a "duplicate" record, and why can duplicates be a problem?

Accepted Answer

The same record appearing more than once, which can bias learning

Answer

A row that is missing all values

Answer

A row written in another language

Answer

A column that has only numbers

Question 6

What is an "outlier" in a dataset?

Accepted Answer

A value far from most others that may be rare or an error

Answer

A value that is missing

Answer

The most common value in the data

Answer

A value that is exactly the average

Question 7

Why might the same information be written inconsistently, like "NYC", "New York", and "new york"?

Accepted Answer

Different people type the same thing in different ways

Answer

The computer changes the words randomly

Answer

It is a secret code

Answer

The data file is broken

Question 8

What does it mean to "standardize" text values during cleaning?

Accepted Answer

Putting values into one consistent format

Answer

Translating text into another language

Answer

Deleting all text columns

Answer

Making text appear bigger

Question 9

In a spreadsheet of data, what do rows and columns usually represent?

Accepted Answer

Rows are examples and columns are features

Answer

Rows are colors and columns are sizes

Answer

Rows and columns are both just decoration

Answer

Rows are features and columns are files

Question 10

What is a "data type", and why does it matter when cleaning data?

Accepted Answer

The kind of value (number, text, date) a column holds

Answer

The color of the data

Answer

How many rows the data has

Answer

The name of the file

Question 11

Why might you convert the text "42" into the number 42 during cleaning?

Accepted Answer

So the computer can do math with it

Answer

So it looks prettier

Answer

So it takes more space

Answer

Text and numbers are always the same

Question 12

What is "normalization" (scaling) of numeric data?

Accepted Answer

Rescaling numbers into a common range like 0 to 1

Answer

Deleting all numbers

Answer

Making all numbers negative

Answer

Rounding numbers to whole numbers only

Question 13

Why is splitting data into "training" and "testing" sets part of preparation?

Accepted Answer

To test the model on data it has not seen for an honest score

Answer

To make the dataset smaller for fun

Answer

Because two files are always better than one

Answer

To hide data from the user

Question 14

What is a "label" in a dataset used for supervised learning?

Accepted Answer

The correct answer attached to an example

Answer

A sticker on the computer

Answer

The name of the dataset file

Answer

A type of error message

Question 15

What problem can mislabeled examples cause for a model?

Accepted Answer

The model learns wrong patterns and makes more mistakes

Answer

Nothing, labels do not matter

Answer

The computer runs faster

Answer

The data file becomes smaller

Question 16

What does it mean to "trim whitespace" from data?

Accepted Answer

Removing extra spaces around or inside a value

Answer

Cutting the data file in half

Answer

Changing letters to numbers

Answer

Adding more spaces for readability

Question 17

Why should you be careful before deleting rows that have missing data?

Accepted Answer

You might lose useful information and shrink the data too much

Answer

Deleting rows always improves accuracy

Answer

Rows can never be deleted by a computer

Answer

Missing data is impossible to detect

Question 18

What is "imputation" in data cleaning?

Accepted Answer

Filling missing values with estimated ones

Answer

Deleting columns

Answer

Doubling every value

Answer

Sorting the data alphabetically

Question 19

When filling missing numbers, why might the median be safer than the mean if there are extreme outliers?

Accepted Answer

The mean can be distorted by extreme values, the median resists them

Answer

The median is always larger

Answer

The median is easier to spell

Answer

They are exactly the same thing

Question 20

What does "consistent units" mean, and why does it matter in a dataset?

Accepted Answer

Using the same measurement unit in a column

Answer

Using the same color for all values

Answer

Having the same number of rows and columns

Answer

Writing all values in capital letters

Question 21

What is a "feature" in machine learning data?

Accepted Answer

One measurable input piece of information about an example

Answer

A special button on a computer

Answer

The final answer of the model

Answer

A type of chart

Question 22

What is "encoding" categorical data, and why is it needed?

Accepted Answer

Turning categories into numbers so models can use them

Answer

Hiding data with a secret password

Answer

Deleting all text columns

Answer

Making the data run faster

Question 23

How can a simple chart, like a bar chart or scatter plot, help during data cleaning?

Accepted Answer

Charts help you spot strange values, gaps, and outliers

Answer

Charts only make the data look nicer

Answer

Charts delete bad data automatically

Answer

Charts are not useful for cleaning

Question 24

What is the risk of having "biased" data, and how does it connect to ethics?

Accepted Answer

Biased data can make models unfair to some groups

Answer

Biased data makes computers faster

Answer

Bias only affects the file size

Answer

Bias has nothing to do with fairness

Question 25

Why should personal or private information be handled carefully when preparing data?

Accepted Answer

To protect people, you should anonymize or remove sensitive details

Answer

Private data is never in datasets

Answer

Private data should always be shared widely

Answer

Privacy does not matter in data work

Question 26

What does "anonymizing" data mean?

Accepted Answer

Removing or disguising details that could identify a person

Answer

Adding more names to the data

Answer

Translating the data into another language

Answer

Making the data load faster

Question 27

What is a "data dictionary" and how does it help with preparation?

Accepted Answer

A document explaining what each column means

Answer

A dictionary of programming words

Answer

A list of deleted rows

Answer

A tool that translates languages

Question 28

Why is it useful to look at the minimum and maximum of a numeric column while cleaning?

Accepted Answer

It reveals impossible or suspicious extreme values

Answer

It makes the data smaller

Answer

Minimum and maximum are never useful

Answer

It changes the values automatically

Question 29

What is "validation" of data, such as checking that ages are between 0 and 120?

Accepted Answer

Checking that values follow expected rules

Answer

Making data look colorful

Answer

Deleting all numbers

Answer

Copying data to a new file

Question 30

What is "feature engineering" with a simple example?

Accepted Answer

Creating new useful features from existing data

Answer

Building physical machines

Answer

Deleting all features

Answer

Renaming the dataset

Test yourself