Maximizing Python Speed with Numpy Vectorization (Part 1)

Why do people say Python is slow? How do you analyze a Python algorithm to find room for improvement?

We will walk you through the steps of how to think about optimizing a time series clustering algorithm using numpy vectorization techniques.

In Part 1 of this series, Sean will explain why numpy is fast and dive into the code that reduces the benchmark from 6 minutes to less than 10 seconds.

0:48 Why is SQL slow for this?
1:45 The essence of the problem
3:03 Agglomerative clustering
4:31 Why list of lists is slow?
5:04 What does contiguous mean?
7:05 How does vectorization help us?

MUSIC

[ShaunakDe]

This is such an under-rated skill with so few resources available to help people get better at doing vectorization! Thanks for putting this out.