Why Setting random seed for reproducibility in NumPy? - Purpose & Use Cases

The Big Idea

What if you could make randomness behave the same way every time you run your code?

The Scenario

Imagine you are running an experiment that involves random numbers, like shuffling a deck of cards or picking random samples from a dataset. You write your code, get some results, but when you run it again, the results change unexpectedly.

The Problem

Without controlling randomness, every run produces different outcomes. This makes it hard to debug, compare results, or share your work with others. Manually trying to recreate the exact random sequence is nearly impossible and very frustrating.

The Solution

By setting a random seed, you tell the computer to start the random number generator from the same point every time. This means your random results become predictable and repeatable, making your experiments reliable and easier to share.

Before vs After

✗ Before

import numpy as np
random_numbers = np.random.rand(5)
print(random_numbers)

✓ After

import numpy as np
np.random.seed(42)
random_numbers = np.random.rand(5)
print(random_numbers)

What It Enables

It enables you to produce the same random results every time, making your data science experiments trustworthy and easy to reproduce.

Real Life Example

A data scientist shares a machine learning model with colleagues. By setting the random seed, everyone gets the same training and testing splits, ensuring consistent evaluation and fair comparison.

Key Takeaways

Randomness without control leads to inconsistent results.

Setting a random seed fixes the starting point of randomness.

This makes experiments repeatable and results reliable.

Practice

1. What is the main purpose of setting a random seed in numpy?

easy

A. To increase the range of random numbers

B. To speed up random number generation

C. To generate only positive random numbers

D. To make random number results repeatable

Solution

1. Step 1: Understand what a random seed does

   Setting a random seed fixes the starting point for random number generation.

2. Step 2: Effect of fixed seed on results

   This means the same sequence of random numbers is produced every time the code runs.

3. Final Answer:

   To make random number results repeatable -> Option D

4. Quick Check:

   Random seed = repeatable results [OK]

Hint: Seed fixes randomness to get same results every run [OK]

Common Mistakes:

• Thinking seed speeds up random generation
• Believing seed changes number range
• Assuming seed makes numbers positive only

2. Which of the following is the correct syntax to set a random seed to 42 in numpy?

easy

A. np.random.seed(42)

B. np.seed.random(42)

C. np.random.set_seed(42)

D. np.random.seed = 42

Solution

1. Step 1: Recall the correct function name and usage

   The correct function to set seed in numpy is np.random.seed().

2. Step 2: Check the argument and syntax

   The seed value is passed as an integer inside the parentheses, like np.random.seed(42).

3. Final Answer:

   np.random.seed(42) -> Option A

4. Quick Check:

   Correct function call = np.random.seed(42) [OK]

Hint: Use np.random.seed(number) exactly [OK]

Common Mistakes:

• Swapping function and module names
• Using incorrect function like set_seed
• Assigning seed instead of calling function

3. What will be the output of the following code?

import numpy as np
np.random.seed(0)
print(np.random.randint(1, 10))
np.random.seed(0)
print(np.random.randint(1, 10))

medium

A. Different numbers both times

B. 5 followed by different number

C. 5 followed by 5

D. Error due to repeated seed setting

Solution

1. Step 1: Understand effect of setting seed before each random call

   Setting the seed to 0 resets the random number generator to the same start point.

2. Step 2: Predict output of np.random.randint(1, 10) after resetting seed

   Both calls produce the same random integer because the seed is reset before each call.

3. Final Answer:

   5 followed by 5 -> Option C

4. Quick Check:

   Same seed = same random number [OK]

Hint: Resetting seed repeats same random number [OK]

Common Mistakes:

• Assuming different numbers after resetting seed
• Thinking repeated seed causes error
• Ignoring seed reset effect

4. Identify the error in this code snippet that tries to set a random seed:

import numpy as np
np.random.seed = 123
print(np.random.rand())

medium

A. Missing import statement

B. Seed is assigned instead of called as a function

C. Seed value must be a float, not integer

D. np.random.rand() is incorrect function

Solution

1. Step 1: Check how seed is set

   The code assigns 123 to np.random.seed instead of calling it as a function.

2. Step 2: Understand consequence of assignment

   This overwrites the seed function with an integer, causing errors on later random calls.

3. Final Answer:

   Seed is assigned instead of called as a function -> Option B

4. Quick Check:

   Use np.random.seed(123), not assignment [OK]

Hint: Call seed() as function, don't assign it [OK]

Common Mistakes:

• Assigning seed instead of calling
• Using wrong random function
• Ignoring import statement

5. You want to generate a reproducible array of 3 random floats between 0 and 1 using numpy. Which code snippet achieves this correctly?

hard

A. np.random.seed(7)
arr = np.random.rand(3)
print(arr)

B. np.random.seed = 7
arr = np.random.rand(3)
print(arr)

C. np.random.set_seed(7)
arr = np.random.rand(3)
print(arr)

D. np.random.seed(7)
arr = np.random.randint(3)
print(arr)

Solution

1. Step 1: Set the random seed correctly

   Use np.random.seed(7) to fix the random sequence.

2. Step 2: Generate 3 random floats between 0 and 1

   np.random.rand(3) creates an array of 3 floats in [0,1).

3. Final Answer:

   np.random.seed(7) arr = np.random.rand(3) print(arr) -> Option A

4. Quick Check:

   Correct seed + rand(3) = reproducible floats [OK]

Hint: Seed with np.random.seed(), use np.random.rand(3) [OK]

Common Mistakes:

• Assigning seed instead of calling
• Using non-existent set_seed function
• Using randint instead of rand for floats