Yes, It is possible to use the NaN variable without NumPy in Python.
You can use float('nan') for the following. In Python 3.5, you can also use math.nan.
>>> a = float('nan')
>>> print(a)
nan
>>> print(a + 2)
nan
>>> a == a
False
>>> import math
>>> math.isnan(a)
True
>>> # Python 3.5+
>>> math.isnan(math.nan)
True