Clipping hot and cold pixels is a very useful technique, but it's important to understand why it's needed.
Most of the images we process are single-channel, representing the intensity of a measured quantity, rather than typical photographs. Users interpret these intensities by referring to a color scale or color bar. For example, if a user sees the color red, they'll consult the color bar to understand that red corresponds to a value of, say, 400. This interpretation is a crucial aspect of the functionality.
When displaying a color bar, Matplotlib assumes that the lowest color corresponds to the lowest value in the image (potentially a "cold" pixel) and the highest color corresponds to the highest value in the image (potentially a "hot" pixel). After hot/cold pixel removal, the color scale can be displayed correctly.
The ability to evaluate the effectiveness of hot/cold pixel removal relies entirely on the proper display of the color bar/color scale.