CSC format (Compressed Sparse Column) in SciPy - Mini Project: Build & Apply

Working with CSC format (Compressed Sparse Column) in SciPy

📖 Scenario: You have a sparse matrix representing connections between users and items in a recommendation system. Sparse matrices save memory by storing only non-zero values.One common way to store sparse matrices is the CSC (Compressed Sparse Column) format, which stores data column-wise efficiently.

🎯 Goal: You will create a sparse matrix in COO format, convert it to CSC format, and then extract the data arrays that represent the CSC structure.

📋 What You'll Learn

Create a COO sparse matrix with given data, row indices, and column indices

Create a variable for the shape of the matrix

Convert the COO matrix to CSC format

Extract the data, indices, and indptr arrays from the CSC matrix

Print the extracted arrays

💡 Why This Matters

🌍 Real World

Sparse matrices are used in recommendation systems, natural language processing, and scientific computing where data is mostly zeros. CSC format helps efficiently store and process such data.

💼 Career

Data scientists and machine learning engineers often work with sparse data. Understanding CSC format helps optimize memory and speed when handling large datasets.

1

Create a COO sparse matrix

Import coo_matrix from scipy.sparse. Create a COO sparse matrix called coo with data = [10, 20, 30, 40], row = [0, 1, 2, 0], and col = [0, 2, 2, 1]. Set the shape to (3, 3).

SciPy

# Import coo_matrix and create the COO sparse matrix coo
# Your code here

Need a hint?

Use coo_matrix((data, (row, col)), shape=(3, 3)) to create the sparse matrix.

2

Set the shape variable

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

SciPy

from scipy.sparse import coo_matrix

data = [10, 20, 30, 40]
row = [0, 1, 2, 0]
col = [0, 2, 2, 1]
coo = coo_matrix((data, (row, col)), shape=(3, 3))

# Create a variable called shape and set it to (3, 3)
# Your code here

Need a hint?

Just assign the tuple (3, 3) to the variable shape.

3

Convert COO to CSC format and extract arrays

Convert the COO matrix coo to CSC format and store it in a variable called csc. Then extract the data, indices, and indptr arrays from csc and store them in variables called data_csc, indices_csc, and indptr_csc respectively.

SciPy

from scipy.sparse import coo_matrix

data = [10, 20, 30, 40]
row = [0, 1, 2, 0]
col = [0, 2, 2, 1]
coo = coo_matrix((data, (row, col)), shape=(3, 3))

shape = (3, 3)

# Convert coo to csc format and extract data, indices, indptr arrays
# Your code here

Need a hint?

Use tocsc() method to convert. Access data, indices, and indptr attributes from the CSC matrix.

4

Print the CSC arrays

Print the variables data_csc, indices_csc, and indptr_csc each on a separate line.

SciPy

from scipy.sparse import coo_matrix

data = [10, 20, 30, 40]
row = [0, 1, 2, 0]
col = [0, 2, 2, 1]
coo = coo_matrix((data, (row, col)), shape=(3, 3))

shape = (3, 3)

csc = coo.tocsc()
data_csc = csc.data
indices_csc = csc.indices
indptr_csc = csc.indptr

# Print data_csc, indices_csc, and indptr_csc each on a new line
# Your code here

Need a hint?

Use three print() statements, one for each variable.