When NumPy is not fast enough - Mini Project: Build & Apply

When NumPy is not fast enough

📖 Scenario: You are working with large numerical datasets using NumPy. Sometimes, NumPy operations can be slow for very large or complex computations. To speed things up, you want to try a faster alternative using numba, a tool that can compile Python code to machine code.This project will guide you through creating a NumPy array, setting a threshold, writing a function with numba to speed up calculations, and comparing the results.

🎯 Goal: Build a small program that creates a NumPy array, sets a threshold value, uses numba to speed up a calculation that counts how many elements are above the threshold, and prints the result.

📋 What You'll Learn

Create a NumPy array with specific values

Set a threshold variable

Write a function using numba to count elements above the threshold

Print the count result

💡 Why This Matters

🌍 Real World

In data science, sometimes NumPy alone is not fast enough for complex or large computations. Using tools like numba can speed up your code by compiling Python functions to machine code.

💼 Career

Knowing how to optimize numerical computations with numba is valuable for data scientists and engineers working with large datasets or performance-critical applications.

1

Create the NumPy array

Import numpy as np and create a NumPy array called data with these exact values: [1, 5, 8, 12, 20, 3, 7].

NumPy

# Import numpy as np and create the array data
# Your code here

Need a hint?

Use np.array([...]) to create the array.

2

Set the threshold value

Create a variable called threshold and set it to the integer 7.

NumPy

import numpy as np

data = np.array([1, 5, 8, 12, 20, 3, 7])

# Create the threshold variable and set it to 7
# Your code here

Need a hint?

Just assign the number 7 to threshold.

3

Write a numba-accelerated function

Import njit from numba. Then write a function called count_above_threshold that takes two arguments: arr and thresh. Use a for loop to count how many elements in arr are greater than thresh. Decorate the function with @njit.

NumPy

import numpy as np
from numba import njit

data = np.array([1, 5, 8, 12, 20, 3, 7])

threshold = 7

# Write the @njit decorated function count_above_threshold(arr, thresh)
# Your code here

Need a hint?

Use @njit above the function. Inside, loop over arr and count elements greater than thresh.

4

Print the count result

Call the function count_above_threshold with data and threshold as arguments. Print the returned count.

NumPy

import numpy as np
from numba import njit

data = np.array([1, 5, 8, 12, 20, 3, 7])

threshold = 7

@njit
def count_above_threshold(arr, thresh):
    count = 0
    for x in arr:
        if x > thresh:
            count += 1
    return count

# Call count_above_threshold with data and threshold, then print the result
# Your code here

Need a hint?

Use print(count_above_threshold(data, threshold)) or assign to a variable first.