First and Foremost, We are Data Scientists
We understand that our analyses inform our customers in making significant investments. We take our responsibility seriously.
We use the tools of model selection, training, validation, and testing to assure and quantify the validity of our analyses by the standards and best practices of the field. Scientific rigor and professional commitment have been built into our platform as a design principal from the outset of our development. From end to end, our platform architecture is designed to provide accountability of feature significance, and traceable provenance for all features—engineered, or otherwise. Our focus on verifiability and transparency enhances the credibility of our work, and provides the validity demanded by the highest professional and academic standards.
For Koan, the quality of the work is a team effort. A high degree of scrutiny, and critical discussion of potential findings is necessary, and engaged. In the design of new models or features, we ensure subject-matter experts are paired with data scientist/developers. In this way, the rigorous interpretation of findings by domain experts can help provide appropriate optimization, selection, training, validation, and testing approaches.