Step 1 creates a small sample DataFrame that will help us illustrate the differences between two and one-variable plotting with pandas. By default, pandas will use each numeric column of the DataFrame to make a new set of bars, lines, KDEs, boxplots, or histograms and use the index as the x values when it is a two-variable plot. One of the exceptions is the scatter plot, which must be explicitly given a single column for the x and y values.

The pandas plot method is very versatile and has a large number of parameters that allow you to customize the result to your liking. For instance, you can set the Figure size, turn the gridlines on and off, set the range of the x and y axis, color the plot, rotate the tick marks, and much more.

You can also use any of the arguments available to the specific matplotlib plotting method. The extra arguments will be collected by the **kwds parameter from the plot method and correctly passed to the underlying matplotlib function. For example, In step 2, we create a bar plot. This means that we can use all of the parameters available in the matplotlib bar function as well as the ones available in the pandas plot method (http://bit.ly/2z2l3rJ).

In step 3, we create a single-variable KDE plot, which creates a density estimate for each numeric column in the DataFrame. Step 4 places all the two-variable plots in the same Figure. Likewise, step 5 places all the one-variable plots together. Each of steps 4 and 5 creates a Figure with three Axes objects. The command plt.subplots(1, 3) creates a Figure with three Axes spread over a single row and three columns. It returns a two-item tuple consisting of the Figure and a one-dimensional NumPy array containing the Axes. The first item of the tuple is unpacked into the variable fig. The second item of the tuple is unpacked into three more variables, one for each Axes. The pandas plot method handily comes with an ax parameter, allowing us to place the result of the plot into a specific Axes in the Figure.