WEEK 6

Question 1

MultiIndex in pandas

Answer 1

allows multiple levels of indexing, making it easier to organize and manipulate hierarchical data.

Question 2

explain unstack and stack

Answer 2

unstack (spread it out): Turns a MultiIndex Series into a DataFrame (makes it wide).

stack (pull it together): turns a DataFrame back into a MultiIndex Series.

Question 3

why do we use multi index concept

Answer 3

it allows us to analyze complex dataset better

Question 4

array

Answer 4

states = [‘California’, ‘Texas’, ‘New York’]
years = [2000, 2010, 2000]

Here, both states and years are arrays (lists of values).

Question 5

tulip

Answer 5

A tuple is like an array, but it groups values together as one item.

(‘California’, 2000) is a tuple acting as a key in this dictionary.

Question 6

What is the purpose of np.concatenate() in NumPy?

Answer 6

It combines multiple arrays into one single array.

Question 7

What is pd.concat() used for?

Answer 7

It combines Series or DataFrames along rows (axis=0) or columns (axis=1).

Question 8

If you pd.concat([ser1, ser2]), what happens?

Answer 8

It stacks ser2 below ser1, combining them into one Series.

Question 9

What happens in an inner join?

Answer 9

Keeps only common columns between DataFrames.

Question 10

What happens in an outer join?

Answer 10

Keeps all columns; missing values are filled with NaN.

Question 11

Does pandas keep duplicate indices when combining DataFrames?

Answer 11

Yes, by default pandas keeps duplicates.

Question 12

What are three ways to handle duplicate indices when joining DataFrames?

Answer 12

• Reset index (ignore_index=True)
• Group using MultiIndex (keys=[‘First’, ‘Second’])
• Use verify_integrity=True to stop if duplicates exist

Question 13

What happens when you use .append() on DataFrames?

Answer 13

It stacks one DataFrame on top of another, keeping the original indices unless reset.

Question 14

What is relational algebra in simple terms?

Answer 14

A set of rules for working with table-based data, used in databases like SQL.

Question 15

What is the purpose of pd.merge() in pandas?

Answer 15

It joins two tables based on a common column, like SQL joins.

Question 16

What are the three types of joins based on data relationships?

Answer 16

One-to-One: Both tables have unique keys.

Many-to-One: One table has repeated keys.

Many-to-Many: Both tables have repeated keys.

Question 17

What does ‘inner’ join do?

Answer 17

Keeps only matching rows.

Question 18

What does ‘outer’ join do?

Answer 18

Keeps all rows, fills missing values with NaN.

Question 19

What does ‘left’ join do?

Answer 19

Keeps all rows from the first table, fills missing from the second.

Question 20

What does ‘right’ join do?

Answer 20

Keeps all rows from the second table, fills missing from the first.

Question 21

What does aggregation mean in pandas?

Answer 21

Summarizing data by applying functions like sum(), mean(), max(), etc.

Question 22

What does a pivot table do?

Answer 22

It groups data by a specific category (like gender or class) and applies aggregation to find totals, averages, etc.

Question 23

Why use pivot tables in pandas?

Answer 23

They make it easy to compare data across categories and summarize large datasets.

Question 24

What helps pandas clean messy text data quickly?

Answer 24

Vectorized string operations, faster than loops.

Question 25

How can you turn text dates into proper date objects?

Answer 25

Use dateutil.parser.

Question 26

What does datetime64 do in NumPy?

Answer 26

Enables fast date-time operations.

Question 27

What can pd.eval() handle?

Answer 27

Arithmetic, comparisons, and boolean logic.

Question 28

Why use df.eval()?

Answer 28

Makes it easier to reference column names directly.

Question 29

Q: How do you reference Python variables inside eval()?

Answer 29

Use @ symbol.

Question 30

What does slicing mean in data analysis?

Answer 30

Selecting a subset of data based on one dimension — like a few rows or one column.

Question 31

What does dicing mean in data analysis?

Answer 31

Selecting a specific subset of data based on multiple dimensions or criteria.

Question 32

What is the key difference between slicing and dicing?

Answer 32

Slicing → 1 dimension (rows or columns) Dicing → multiple dimensions (rows and columns or filters)

Question 33

Why do we use slicing or dicing?

Answer 33

To extract just the data we need for analysis or to focus on specific groups or time periods.

Question 34

what is more expensive: dicing or slicing?

Answer 34

dicing

Question 35

What is a full outer join?

Answer 35

Keeps everything from both DataFrames, filling missing spots with NaN.

Question 36

What does NaN stand for in pandas?

Answer 36

Not a Number

Question 37

What is the use of .isna() in pandas?

Answer 37

Helps detect missing values.

Question 38

What is considered bad data?

Answer 38

Empty cells, wrong formats, wrong data, or duplicates.

Question 39

What is one way to fix wrong formats in pandas?

Answer 39

You can drop the incorrect data or convert it to the correct format.

Question 40

Why is removing duplicates important?

Answer 40

To clean data and avoid wrong analysis or double counting.