COO format (Coordinate) in SciPy - Mini Project: Build & Apply

Working with COO format sparse matrices using scipy

📖 Scenario: Imagine you are working with a large dataset representing connections between users in a social network. Most users are not directly connected, so the data is mostly zeros. To save memory and speed up calculations, you want to use a sparse matrix format called COO (Coordinate format).

🎯 Goal: You will create a sparse matrix in COO format using scipy, configure its shape, add data points, and finally display the matrix in dense form to see the connections.

📋 What You'll Learn

Create three lists: row, col, and data with exact values

Create a variable shape as a tuple for matrix dimensions

Use scipy.sparse.coo_matrix with data, (row, col), and shape

Print the dense form of the COO matrix using .toarray()

💡 Why This Matters

🌍 Real World

Sparse matrices are used in social networks, recommendation systems, and scientific computing where most data points are zero.

💼 Career

Understanding sparse matrix formats helps in data science and machine learning jobs that handle large, sparse datasets efficiently.

1

Create the COO matrix data lists

Create three lists called row, col, and data with these exact values: row = [0, 1, 2, 0], col = [1, 2, 0, 2], and data = [4, 5, 7, 9].

SciPy

# Create lists row, col, and data with the specified values
# Your code here

Need a hint?

Remember to create three separate lists named exactly row, col, and data with the values given.

2

Set the shape of the sparse matrix

Create a variable called shape and set it to the tuple (3, 3) to define the matrix size.

SciPy

row = [0, 1, 2, 0]
col = [1, 2, 0, 2]
data = [4, 5, 7, 9]
# Create a tuple called shape with value (3, 3)
# Your code here

Need a hint?

The shape tuple defines the number of rows and columns in the matrix.

3

Create the COO sparse matrix

Import coo_matrix from scipy.sparse and create a variable called matrix by passing data, (row, col), and shape to coo_matrix.

SciPy

row = [0, 1, 2, 0]
col = [1, 2, 0, 2]
data = [4, 5, 7, 9]
shape = (3, 3)
# Import coo_matrix and create matrix using data, (row, col), and shape
# Your code here

Need a hint?

Use the syntax coo_matrix((data, (row, col)), shape=shape) to create the sparse matrix.

4

Display the dense matrix

Print the dense form of the COO matrix by calling matrix.toarray() inside a print() statement.

SciPy

from scipy.sparse import coo_matrix

row = [0, 1, 2, 0]
col = [1, 2, 0, 2]
data = [4, 5, 7, 9]
shape = (3, 3)
matrix = coo_matrix((data, (row, col)), shape=shape)
# Print the dense matrix using matrix.toarray()
# Your code here

Need a hint?

The printed output should be a 3x3 matrix showing the non-zero values at the correct positions.