Exploratory data analysis (EDA) is an approach to data analysis used to investigate sets of data, summarize their characteristics, and figure out how to best work with data to get answers while providing a visual to help businesses, scientists, researchers, and analysts learn more from that data. Exploratory data analysis makes it easier to find patterns and anomalies in data, and it can be used to determine what the data reveals beyond modeling. It's useful as a step in creating sophisticated data models and analysis. EDA tools include clustering/dimension reduction techniques to create graphs, K-means clustering, and predictive modeling, including linear regression. There are four main types of exploratory data analysis, including univariate non-graphical, univariate graphical, multivariate nongraphical, and multivariate graphical. All of these types describe the data, but graphical exploratory data analysis provides a more complete picture created by the data.