Thesis for the degree of Master of Science Balancing Cost and Precision of Approximate Type Inference in Python

In dynamically typed programming languages, values have types, but vari-ables and other constructs in the source code do not. A variable can thus refer to any type at runtime, and in general, it is not possible to statically determine these types. Nonetheless, it is often useful to have some information on the types of variables, for example for interactive tools such as editors and IDEs (integrated development environments). This thesis presents a method for approximate type inference for the dy-namically typed language Python. It focuses on balancing precision and cost to find a method that is fast enough to be used in interactive tools while still yielding useful type information. To this end, several different variants of the basic method were developed, and their speed and precision were evaluate