Exploratory Data Analysis (EDA)
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
The practice of summarizing and visualizing a dataset to understand its structure, spot quality problems, and surface patterns before formal modeling.
What is Exploratory Data Analysis (EDA)?
Exploratory data analysis (EDA) is the practice of summarizing and visualizing a dataset to understand its structure, spot quality problems, and surface patterns before formal modeling begins.
The statistician John Tukey named the approach in his 1977 book Exploratory Data Analysis, arguing that analysts should let data suggest hypotheses instead of only testing the ones they arrived with. Every data science project still starts here: build a model on a dataset you haven’t explored, and its surprises become the model’s errors.
Types of EDA Techniques
- Univariate analysis examines one variable at a time, with histograms, box plots, and summary statistics that show each variable’s distribution and outliers.
- Bivariate analysis examines pairs, with scatter plots and correlations that reveal how two variables move together.
- Multivariate analysis examines many variables at once, with pair plots and heatmaps that expose interactions no single chart shows.
In practice this work happens in a Jupyter notebook, with a library such as pandas computing the summaries and a plotting library drawing the charts. The notebook format suits EDA because the questions arrive one at a time, and each answer decides what to look at next.
Example of Exploratory Data Analysis
An analyst receives a customer churn dataset and explores it before any modeling. Histograms show monthly charges split into two distinct groups, a missing-value check finds a block of empty income fields, and a plot of tenure against churn shows cancellations cluster in the first three months.
One finding earns its keep: hundreds of accounts show zero usage, and they turn out to be a logging bug rather than real customers. Dropping them before feature engineering prevents the model from learning a pattern that doesn’t exist.
Related AI terms: Data Science · Data Analytics · Data Quality · Multivariate Analysis · Feature Engineering
Did you like the Exploratory Data Analysis (EDA) gist?
Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.
Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About