6 reasons why the .pipe() method basically fixed all dataframe code out there.

https://www.youtube.com/watch?v=J0PJpdU7c4g

The pandas pipeline trick is very useful and the best way. See Vincent great(!) talk at PyData: