COMP 2304 · Data Science: Python for Analysis

Led by Downeyian Computational Thinking Simulacrum

Python for data analysis — from foundations through NumPy and pandas to exploratory data analysis. Based on the works of Allen Downey.

1. Module 1

   Python Foundations

   Led by Downeyian Computational Thinking Simulacrum

   The question

   A variable stores a value so you can reuse it. A loop repeats an operation. A function names a procedure. These are tools. What are the eight tools a data scientist must know to do anything useful with Python?

   Outcome

   The student can write Python functions and use comprehensions for data transformation.

   Sub-units

   1. ○ 1.1 Write Three Functions Open this unit → ✓ ↻
   2. ○ 1.2 Loops and Comprehensions Open this unit → ✓ ↻
   Begin the course → Print certificate of completion
2. Module 2

   NumPy: Arrays and Vectorised Computation

   Led by Downeyian Computational Thinking Simulacrum

   The question

   Every ML library in Python uses NumPy arrays as its core data structure. Vectorised operations are 10-100x faster than Python loops. Why — and what does broadcasting mean?

   Outcome

   The student can manipulate NumPy arrays and explain why vectorisation outperforms loops.

   Sub-units

   1. ○ 2.1 Vectorised vs Loop Open this unit → ✓ ↻
   Open this module → Print certificate of completion
3. Module 3

   pandas: DataFrames for Data Analysis

   Led by Downeyian Computational Thinking Simulacrum

   The question

   Eight operations cover 80% of data wrangling: import, inspect, select, filter, handle missing values, transform, aggregate, merge. What are they and why does each matter?

   Outcome

   The student can execute a complete data wrangling pipeline in pandas.

   Sub-units

   1. ○ 3.1 Data Wrangling Pipeline Open this unit → ✓ ↻
   Open this module → Print certificate of completion
4. Module 4

   Visualisation: Matplotlib and Seaborn

   Led by Downeyian Computational Thinking Simulacrum

   The question

   A visualisation is an argument. The choice of chart is a choice about what claim you are making. What are the right charts for distributions, relationships, and comparisons — and what makes a chart misleading?

   Outcome

   The student can produce histograms, scatter plots, box plots, and heatmaps with interpretation.

   Sub-units

   1. ○ 4.1 Four Charts Open this unit → ✓ ↻
   Open this module → Print certificate of completion
5. Module 5

   Exploratory Data Analysis in Practice

   Led by Downeyian Computational Thinking Simulacrum

   The question

   EDA is what happens before modelling. You look at your data, form hypotheses, and discover things not in the problem statement. What does a complete EDA look like — and why is building a model on data you have not looked at the most common modelling error?

   Outcome

   The student can conduct and write up a complete EDA on a real dataset.

   Sub-units

   1. ○ 5.1 Final EDA Report Open this unit → ✓ ↻
   Open this module → Print certificate of completion