Immutable data structures reduce opportunities for error and promote the design of pure functions, offering programs that are easier to reason about and maintain. While Pandas is used in many domains where such benefits are highly desirable, there is no way to enforce immutability in Pandas.

StaticFrame offers an alternative to Pandas, leveraging familiar Pandas interfaces and operations while enforcing immutability and providing more narrow and consistent interfaces. StaticFrame depends only on NumPy, and offers comparable performance in many areas.

In this talk I will introduce the StaticFrame library. I will demonstrate how it accomplishes immutability while still permitting flexible assignment-to-copy statements and grow-only, column-based operations.

Additional design choices and features will be discussed, including more narrow and uniform operations and interfaces, the strict prohibition of duplicate index labels, embracing the full range of NumPy types, and a family of axis and value iterators that support function application with built-in concurrency.