Module 3 · Unit 9 · 9 tutorials · Week 7
Exploratory Data Analysis
Univariate, bivariate and correlational analysis. Bringing statistics and visualisation together to discover what a dataset actually contains.
Overview
Exploratory data analysis is the process of getting to know a dataset before you analyse it formally. It is iterative, visual and unhurried: you examine one variable at a time, then pairs of variables, then the relationships among several, adjusting your questions as the data reveals what it contains.
This unit teaches the workflow and the discipline that goes with it — including the honesty required to distinguish what the data shows from what you would like it to show.
Learning outcomes
By the end of this unit you will be able to
- Conduct a structured exploratory analysis of an unfamiliar dataset.
- Examine single variables numerically and visually.
- Analyse relationships between pairs of variables.
- Interpret correlation coefficients responsibly.
- Articulate what patterns do and do not imply.
Tutorial sequence
EDA Overview
The purpose, workflow and mindset of exploratory analysis.
t-02Univariate Analysis
Examining one variable at a time numerically and visually.
t-03Bivariate Analysis
Relationships between two variables and how to test for them.
t-04Correlation Analysis
Correlation coefficients, their interpretation and their limits.
t-05Pattern Discovery
Finding structure worth pursuing and avoiding false patterns.
t-06Handling Outliers in EDA
Detecting unusual observations and deciding how to treat them.
t-07EDA Workflow
A reproducible, systematic process for exploring new data.
t-08Practical Lab
A full EDA on a real dataset, from loading to interpretation.
t-09Unit Summary
Consolidated key points and a forward look to Module 4.
Key concepts
- Exploratory data analysis
- An iterative, largely visual process of examining a dataset to understand its structure before formal analysis.
- Univariate analysis
- Examining a single variable in isolation — its distribution, centre, spread and unusual values.
- Bivariate analysis
- Examining the relationship between two variables, using both statistics and plots.
- Correlation coefficient
- A number between −1 and 1 measuring the strength and direction of a linear relationship, but not implying causation.
- Confounding
- A third variable that influences both of the variables being examined, creating an apparent relationship that is not causal.
Assessment alignment
Directly assessed in Assignment 3
Assignment 3 (Exploratory Data Analysis) requires you to conduct a full EDA on a real-world dataset and to interpret what you find. The workflow and discipline taught here are the substance of that assignment.