Empirical Software Solutions, LLC

Empirical is a programming language designed specifically for time-series analysis, developed by Empirical Software Solutions, LLC. It introduces statically typed dataframes as a core feature, enabling users to perform sophisticated time-series computations with built-in type safety. The language is intended for users who require reliable and efficient manipulation of time-stamped data, such as those working with financial trades, quotes, or events.

A distinguishing aspect of Empirical is its integration of dataframes as native values, allowing for direct loading and manipulation of tabular data within the language. Users can load CSV files as dataframes and perform operations such as filtering, selecting, joining, and aggregating data with concise syntax. The language supports integrated queries and table syntax, making it possible to express complex data transformations and analyses, such as computing the volume-weighted average price (VWAP) or aggregating metrics over custom time intervals. Time-series joins are supported, including as-of joins and nearest joins within specified windows, enabling alignment of disparate time-stamped datasets.

Empirical enforces static typing throughout, with the compiler checking code correctness before execution. This static type system helps prevent common programming errors, particularly in long-running or data-intensive scripts. Type inference is available, allowing dataframe types to be deduced from external sources like file paths at compile time. The language also supports user-defined functions, generic functions, and metaprogramming, facilitating advanced analysis workflows while maintaining type safety.

All Empirical code is compiled to a register-based bytecode for execution on the language's Vector Virtual Machine (VVM), which powers both an interactive REPL and compile-time function evaluation. The platform provides an online interface for trying the language, as well as downloads, tutorials, and source access.