Exploratory Data Analysis (EDA) is literally what it is - Exploring and Analyzing the Data. It is a quite broad term and includes many steps and operations, depending on the problem and data.
Although many people consider it as a separate step, I consider it as a component of the main motive, that is, Data Preparation.
EDA refers to the critical process of performing initial investigations on data so as to discover patterns, spot anomalies, to test hypothesis and check assumptions with the help of summary statistics and graphical visualizations.
This not only helps in understanding the Data we have but also helps in getting the quality of the data right by doing the necessary feature engineering which follows EDA. 
The most common Python libraries used for EDA is Pandas and for the visualization part: Matplotlib (the most popular due to its low-level code and flexibility), Seaborn (sort of a wrapper over Matplotlib and offers quick visualizations in less code) and Plotly (Comes in handy when you want interactive plots).