Engineering the Black-Box Meta Model of Data Exploration

摘要

With an increasing amount and diversity of available data, data exploration is becoming critical for many businesses to create insights for business innovation. From an analysis and design perspective, however, data exploration is still dominated by IT-oriented modeling concepts that make it difficult to engage business users because they are not used to think in terms of (even conceptual) data structures, but rather in terms of business questions, intended insights, decision context, information quality, etc. This study elaborates requirements for conceptualizing data exploration from a business perspective, discusses to what extent existing business-oriented conceptualizations fulfil such requirements, and consolidates promising modeling concepts into a meta model proposal that links purpose, context, domain knowledge, exploration history, business question, available data, user, decision, business insight, presentation and non-user stakeholder as key concepts. The meta model is demonstrated by instantiating it to conceptualize five exemplary data exploration use cases, and first evaluative evidence is presented. The paper closes with a discussion how to transform the proposed meta model into an innovative data exploration analysis and design tool.