np.random.uniform() draws random samples from a uniform distribution over a specified [low, high) range, generalizing np.random.random()'s fixed [0, 1) range to any interval.

Syntax

np.random.uniform(low=0.0, high=1.0, size=None)

Deep Dive Course

Every value between low and high is equally likely to be sampled, with no clustering around any particular point, unlike a normal distribution's characteristic bell-curve concentration around its mean. uniform(low, high, size) is the direct way to simulate any bounded, equally-likely-everywhere quantity, such as a random starting position within a fixed range, or generating test data spread evenly across a known interval.

1Understanding np.random.uniform()

💡

For a custom range, prefer np.random.uniform(low, high, size) directly over manually scaling np.random.random()'s [0, 1) output — it's clearer and avoids getting the scale-and-shift arithmetic wrong.

editor.html

import numpy as np

np.random.seed(0)
print(np.random.uniform(10, 20, 3))

localhost:3000

2Practical Example

Here is a real-world application of np.random.uniform() showing how it is used in production NumPy code.

editor.html

import numpy as np

np.random.seed(0)
temperatures = np.random.uniform(-5, 5, 5)
print(np.round(temperatures, 2))

localhost:3000

3Best Practices

Follow these guidelines when working with np.random.uniform():

1. Use np.random.uniform(low, high, size) directly for a custom range, instead of manually rescaling random()'s [0, 1) output

2. Choose uniform() specifically when every value in a range is genuinely equally likely, and normal() instead when values should cluster around a typical value

3. Seed explicitly for reproducible test data relying on uniform()'s output

⚠️

Tip: For a custom range, prefer np.random.uniform(low, high, size) directly over manually scaling np.random.random()'s [0, 1) output — it's clearer and avoids getting the scale-and-shift arithmetic wrong.

editor.html

import numpy as np

np.random.seed(0)
print(np.random.uniform(10, 20, 3))

localhost:3000

Examples

Example 01Basic Usage

import numpy as np

np.random.seed(0)
print(np.random.uniform(10, 20, 3))

Example 02Advanced Example

import numpy as np

np.random.seed(0)
temperatures = np.random.uniform(-5, 5, 5)
print(np.round(temperatures, 2))

Best Practices

Use np.random.uniform(low, high, size) directly for a custom range, instead of manually rescaling random()'s [0, 1) output
Choose uniform() specifically when every value in a range is genuinely equally likely, and normal() instead when values should cluster around a typical value
Seed explicitly for reproducible test data relying on uniform()'s output

Interview Question

Why does np.random.uniform(10, 20, size) never produce a value clustered near 15 more often than near 10 or 20?▼

Hint: Think about what 'uniform' specifically means, in contrast to a normal distribution.

A uniform distribution assigns exactly equal probability density to every point within its range, by definition — there's no preferred or more-likely value anywhere between low and high, unlike a normal distribution, which is specifically shaped to concentrate probability around its mean and taper off toward the extremes. Over many samples, uniform() values spread out roughly evenly across the entire range rather than clustering anywhere in particular, which is the defining statistical property that distinguishes 'uniform' from 'normal' or other non-uniform distributions.

Exercises

MediumPractice using np.random.uniform() in a real scenario.

View Solution

import numpy as np

np.random.seed(0)
print(np.random.uniform(10, 20, 3))

Frequently Asked Questions

Why does np.random.uniform(10, 20, size) never produce a value clustered near 15 more often than near 10 or 20?▼

Related Functions

np-random-randomnp-random-normalnp-random-randint