🖥️ Week 04 Lecture

From Loops to Vectorisation: NumPy and Pandas

Author

Dr Jon Cardoso-Silva

Published

27 October 2025

📍 Logistics

Time and Location: Thursday, 23 October 2025, 16:00 - 18:00, CLM 5.02

You’ve just submitted the 📝 Week 04 Practice, your first complete data science workflow using pure Python. Today’s lecture revisits the loops you had to write for that exercise. Then we’ll see how to scale that same logic to larger problems using numpy and pandas.

Note: Today’s lecture covered Sections 1-6 of the demonstration notebook (loading data, loops, NumPy vectorisation, Pandas DataFrames, and filtering). Sections 7-8 (custom functions and .apply()) will be introduced in Week 05.

📋 Preparation

Submit the 📝 W04 Practice by today at 12:00 (noon)
Review your loop logic for detecting hot days and counting streaks
Reflect on where you used for loops and if/else statements
Individual feedback on your submission will arrive within a week

🗣️ Lecture Overview

Part 1: W04 Practice Solution (40 min)

Walkthrough of the heatwave detection problem
Dynamic state management in loops: building lists, accumulating counts, tracking streaks
Scaling question: larger datasets or more complex conditions?

Part 2: Vectorisation with NumPy and Pandas (45 min)

NumPy arrays for simple numerical operations
Nested np.where() for complex conditions (ugly!)
Pandas DataFrames with custom functions (cleaner)
Comparing all three approaches

Part 3: Choosing Tools & Lab Preview (15 min)

Decision framework: loops, NumPy, or Pandas?
Tomorrow’s pair programming lab
Mini-Project 1 (released tomorrow)

📓 Lecture Materials

Today’s lecture uses Jupyter Notebooks instead of slides. All materials will be available in your Nuvolos workspace under the week04/ folder, or you can download them directly below.

Lecture Demonstration Notebook

This notebook walks through your W04 Practice solution and introduces NumPy/Pandas vectorisation. It includes Personal Reflection cells where you can take notes during the lecture.