Before the 1970s, statisticians focused heavily on Confirmatory Data Analysis (hypothesis testing). But John Tukey, the father of modern data science, argued that you cannot confirm what you have not yet discovered. He championed , and graphical methods are the sword and shield of EDA.
, the father of Exploratory Data Analysis (EDA), famously championed the use of graphics to find patterns, anomalies, and relationships that statistical tests might miss. A numerical summary can tell you the average, but a graphical method can show you the distribution's shape, the presence of outliers, and the modality (whether data has one peak or multiple peaks). graphical methods for data analysis pdf
Univariate analysis involves examining a single variable. It is the foundation of data understanding. Most comprehensive guides and PDF textbooks on this subject will begin with these fundamental plots. , the father of Exploratory Data Analysis (EDA),
Once an analyst understands individual variables, the next step is exploring relationships between two or more variables. This is where predictive modeling often begins. It is the foundation of data understanding