Framework Development for Independent Usability Audit of Machine Learning Software Products

This bachelor thesis examines the Development for Independent Usability Audit of Machine Learning Software Products, particularly those without a graphical interface. The created framework allows developers to assess the usability of their products and identify potential areas for improvement. The framework includes the classification of developers into personality types (Experimenters, Theorists, Pragmatics), a survey to determine the importance of values and assess the quality of their implementation, and the application of cognitive walkthrough methods for a deeper usability analysis. The thesis discusses the results of the surveys and data visualization, and provides recommendations for product improvements. The conclusion considers future development prospects for the framework, including the integration of artificial intelligence tools and automated analysis methods.