Numbers
Notes for Python programmers:
The Wolfram Client Library for Python natively supports many Python numeric types, including NumPy's integer, float16, etc.
The Wolfram Language by default does exact computation whenever it can:
Notes for Java programmers:
Rational numbers are represented symbolically as reduced fractions in the Wolfram Language. Java does not have a built-in rational type, instead always returning numerical approximations by default.
Notes for Python programmers:
Rational numbers are represented symbolically as reduced fractions in the Wolfram Language. Similar functionality in Python requires importing the fractions module.
Use N to get (potentially faster) numerical results:
The Wolfram Language can handle numbers of any precision:
The language automatically tracks precision of results.
Use ` to explicitly indicate the precision to assume in a number:
Notes for Java programmers:
Java code must explicitly use the BigInteger and BigDecimal types for arbitrary-precision integers and decimals.
Notes for Python programmers:
Numbers in the Wolfram Language can become arbitrarily large like Python integers. Floats in Python require importing a third-party library like mpmath.
I represents 
for complex numbers:
Notes for Java programmers:
Java has no internal representation for the imaginary number i. To compute with complex numbers in Java, you must import or create a package.
Notes for Python programmers:
The Wolfram Language provides a number of different stylized forms, such as ⅈ, ⅉ and 
for an imaginary number, and similarly for other constants, to maximize clarity. Additionally, the Wolfram Language will automatically infer whether or not a number is real or complex upon execution. Python by default uses the symbol j to represent an imaginary number, and in general operations will not necessarily work on them.
Matrices are lists of lists:
SparseArray gives sparse arrays.
Notes for Java programmers:
Sparse arrays are not provided by default in Java, typically requiring constructs from third-party libraries.
Notes for Python programmers:
In the Wolfram Language, it is easy to move between arrays and matrices using functions like MatrixForm. In Python, sparse arrays are not provided by default and require third-party libraries. Additionally, they are different classes, which will cause issues when moving between matrices in SciPy and arrays in NumPy.