Python and data libraries quiz

Question 1

In NumPy, what is the key difference between a Python list and a NumPy array (ndarray) for numerical computation?

Accepted Answer

NumPy arrays store one fixed type in contiguous memory and support fast vectorized C operations

Answer

Python lists are always faster than NumPy arrays for math

Answer

NumPy arrays can only hold strings

Answer

There is no real difference; both behave identically

Question 2

What does NumPy 'broadcasting' allow you to do?

Accepted Answer

Perform operations between different-shaped arrays by virtually stretching the smaller one

Answer

Broadcast data over a network

Answer

Force all arrays to be 1-dimensional

Answer

Convert arrays into Python lists automatically

Question 3

In pandas, what is the difference between a Series and a DataFrame?

Accepted Answer

A Series is 1D labeled data; a DataFrame is a 2D table of columns

Answer

A Series is 2D and a DataFrame is 1D

Answer

They are identical objects

Answer

A DataFrame can hold only numbers, a Series only strings

Question 4

What is the main difference between pandas '.loc[]' and '.iloc[]'?

Accepted Answer

.loc is label-based and .iloc is integer-position-based

Answer

Both select only by integer position

Answer

Both select only by label

Answer

'.loc' is for columns and '.iloc' is for rows only

Question 5

Why does NumPy use vectorization instead of explicit Python loops for array math?

Accepted Answer

It runs the loop in compiled C code instead of the slow Python interpreter

Answer

It uses the GPU automatically for all operations

Answer

It reduces the array to a single value

Answer

It makes the code less readable but uses less memory only

Question 6

In pandas, what does df.groupby('col').mean() compute?

Accepted Answer

It groups rows by unique values of 'col' and averages numeric columns per group

Answer

It computes one overall mean ignoring 'col'

Answer

It sorts the DataFrame alphabetically

Answer

It deletes the column 'col'

Question 7

What is the purpose of the matplotlib function plt.show()?

Accepted Answer

It renders and displays the constructed figure

Answer

It saves the figure to disk as PNG

Answer

It clears all current plots

Answer

It installs matplotlib

Question 8

What does the shape attribute of a NumPy array tell you?

Accepted Answer

A tuple giving the number of elements along each dimension

Answer

The total number of elements only

Answer

The data type of the array

Answer

The memory address of the array

Question 9

In pandas, how do NaN values typically arise and what does df.dropna() do?

Accepted Answer

NaN marks missing data; dropna() removes rows/columns containing NaN

Answer

NaN means zero; dropna() replaces it with 0

Answer

NaN is a string; dropna() converts it to text

Answer

dropna() fills NaN with the column mean

Question 10

What is the difference between df.fillna(0) and df.dropna() for handling missing values?

Accepted Answer

fillna replaces NaN with a value; dropna removes rows/columns with NaN

Answer

Both delete missing rows

Answer

Both replace NaN with the mean

Answer

fillna removes rows and dropna fills them

Question 11

What does np.arange(0, 10, 2) produce?

Accepted Answer

[0 2 4 6 8]

Answer

[0 2 4 6 8 10]

Answer

[2 4 6 8 10]

Answer

[0 1 2 3 4 5 6 7 8 9]

Question 12

What is the difference between np.zeros((2,3)) and np.ones((2,3))?

Accepted Answer

zeros fills with 0.0, ones fills with 1.0; both are 2x3

Answer

Both fill with 0.0

Answer

zeros makes a 3x2 array and ones a 2x3

Answer

ones creates random values

Question 13

Why is boolean (mask) indexing useful in NumPy and pandas?

Accepted Answer

It filters elements/rows using a True/False condition array

Answer

It sorts the array by value

Answer

It converts numbers to booleans permanently

Answer

It deletes the entire array

Question 14

What does df.head() do by default, and why is it useful?

Accepted Answer

Returns the first 5 rows by default for quick inspection

Answer

Returns the last 5 rows

Answer

Returns column names only

Answer

Sorts and returns all rows

Question 15

What information does df.describe() provide?

Accepted Answer

Summary statistics (count, mean, std, min, quartiles, max) of numeric columns

Answer

Only the column data types

Answer

The number of missing values per column

Answer

A correlation matrix

Question 16

What is the difference between np.array reshaping with reshape(3,4) versus a view and a copy?

Accepted Answer

reshape rearranges data into a new shape, returning a view when possible (else a copy)

Answer

reshape always duplicates all the data

Answer

reshape changes the data values

Answer

reshape only works on 1D arrays

Question 17

In pandas, what does the axis parameter control in operations like df.sum(axis=0) vs df.sum(axis=1)?

Accepted Answer

axis=0 operates down rows (per column); axis=1 operates across columns (per row)

Answer

axis=0 sums rows and axis=1 sums nothing

Answer

Both sum the entire table

Answer

axis=1 sums each column

Question 18

What does the pandas method df['col'].value_counts() return?

Accepted Answer

Counts of each unique value, sorted by frequency descending

Answer

The sum of the column

Answer

The unique values only, unsorted

Answer

The mean per category

Question 19

What is the purpose of np.dot(A, B) or the @ operator for 2D arrays?

Accepted Answer

Matrix multiplication following linear-algebra rules

Answer

Element-wise multiplication of A and B

Answer

Concatenation of A and B

Answer

Element-wise addition

Question 20

How does element-wise multiplication (A * B) differ from matrix multiplication (A @ B) in NumPy?

Accepted Answer

* multiplies corresponding elements; @ does row-by-column matrix multiplication

Answer

Both do matrix multiplication

Answer

Both do element-wise multiplication

Answer

* is matrix multiplication and @ is element-wise

Question 21

What is a pandas Index, and why is it important?

Accepted Answer

A set of labels for rows/columns enabling lookup and automatic alignment

Answer

A list of column data types

Answer

The number of rows in the DataFrame

Answer

A copy of the first column only

Question 22

When you do arithmetic between two pandas Series with different indexes, what happens?

Accepted Answer

They align by index label; unmatched labels become NaN

Answer

They add positionally ignoring labels

Answer

It raises an error

Answer

All values become 0

Question 23

What does np.random.seed(42) accomplish before generating random numbers?

Accepted Answer

It makes the random sequence reproducible across runs

Answer

It generates exactly 42 random numbers

Answer

It produces truly random numbers from hardware

Answer

It clears all random state permanently

Question 24

What is the difference between a histogram and a scatter plot for data visualization?

Accepted Answer

Histogram shows one variable's distribution via bins; scatter shows two variables' relationship

Answer

Both show the relationship between two variables

Answer

Histogram needs two variables, scatter needs one

Answer

They are the same chart with different names

Question 25

What does df.merge(other, on='id', how='inner') do?

Accepted Answer

Combines rows where 'id' matches in both DataFrames (inner join)

Answer

Stacks the DataFrames vertically

Answer

Keeps all rows from both even when unmatched

Answer

Deletes the 'id' column

Question 26

What is the difference between how='inner' and how='left' in a pandas merge?

Accepted Answer

inner keeps only matched keys; left keeps all left rows, filling NaN for unmatched

Answer

Both keep all rows from both tables

Answer

left keeps only matched keys

Answer

inner keeps all rows from the right table

Question 27

What does the apply() method do in pandas, e.g. df['x'].apply(func)?

Accepted Answer

It applies a function to each element (or row/column) and returns the results

Answer

It deletes elements matching a condition

Answer

It only works on integers

Answer

It sorts the Series

Question 28

Why is data normalization or scaling (e.g., min-max scaling) often applied before machine learning?

Accepted Answer

So features share a comparable range and no single feature dominates due to scale

Answer

To increase the number of data points

Answer

Because algorithms cannot read decimals

Answer

To remove all missing values

Question 29

What does np.mean(a, axis=0) compute for a 2D array a?

Accepted Answer

The mean of each column (averaging down the rows)

Answer

The mean of each row

Answer

The single mean of all elements

Answer

The maximum of each column

Question 30

What is the difference between pd.concat and pd.merge in pandas?

Accepted Answer

concat stacks along an axis; merge joins on matching key columns

Answer

Both join on key columns

Answer

Both only stack rows

Answer

concat joins on keys and merge stacks rows

Test yourself