Mastering pandas - Second Edition

User-defined styling options

Apart from the in-built functions, pandas provides us with the option to write our own functions to be used for styling. Let's write a function to change the background color of negative values to red:

 def color_negative(val): 
    color = 'red' if val < 0 else 'green' 
    return 'background-color: %s' % color

Such functions can be applied as a styling option through the apply() and applymap methods of pandas. The applymap method applies the function elementwise. The apply() method can be used to apply styles either row-wise or column-wise by setting the axis parameter to 1 or 0. Setting the axis to None applies the function table-wise. Here, our intended operation is elementwise. Let's use applymap:

rand_df.style.applymap(color_negative) ...