Foundations & the Data-Science Workflow
Exploratory Data Analysis
Exploratory data analysis, or EDA, is the open-minded first look at a dataset where you summarise and plot the data to understand its shape, spot oddities, and form hunches — before committing to any formal test or model. The goal is to let the data surprise you.
Typical EDA includes checking how many values are missing, drawing histograms and scatter plots (散布圖), and looking for outliers (離群值) or impossible values like an age of 200. It is where most real understanding (and most data-quality disasters) are caught. The discipline to remember is that patterns you discover during EDA are hypotheses to test on fresh data, not conclusions — it is easy to fool yourself by hunting through enough charts.
Also called