Design patterns to improve Usability of Text Analytic Tools for Digital Humanists

Abstract

Usability of text analytics tools is still a big issue. This paper presents design patterns for tooling for text analytics for Digital Humanities that are specially suited for users who don’t know how to program and therefore enhance usability especially for this target group. The first pattern evolves around availability of the system as a web API. Further we propose a clear and easy workflow to handle data formats because combining different NLP tools often requires significant and complicated work with data formats. Next we present a format to define workflows or NLP pipelines that are only possible because of the aforementioned principles. We suggest to enable clear workflows between programmers and non-programmers to make clear which tasks can be done with and without programming knowledge. Last we propose a system to allow for reuse of code.