Master's Thesis Matti Maier

Abstract

Enterprise metrics in business accounting are based on a data model in one or more business information systems, while also consisting of multiple influencing variables. The calculation of a metric is often performed on a large data basis, for example on order receipts, accounting information and recipes. When evaluating the calculation results on a large model, the single variable’s impact on the metric is elusive. A simulation of the calculation with changed influencing variables will empower the user to recognize the impact of each influencing variable on all defined enterprise metrics and thus he/she can better react to environmental changes. The work in hand tries to describe an interactive simulation of the changed influencing variable’s impact on the metrics and it tries to answer the question how to visualize the changes and their impact in near real-time.

At first the business and technical requirements for a solution are described, followed by a literature research for validated solutions to the problem at hand and the development of a solution approach to the identified requirements. On the basis of these findings a solution concept is developed. To verify the approach, a prototypical implementation is created. An evaluation with the objective to build an executable web application to interactively simulate and visualize the impact of data changes to enterprise metrics completes the thesis.

Problem

• Calculation of metrics based on certain data and other metrics
• Change of data -> (cascading) change of metric(s) -> change of visualizations -> user perceive change
• Intransparent propagation of change
  • Which metrics are affected, how big is the impact, "positive" or "negativ" impact (based on a certain interpretation of a metric), how to visualize changes?
  • Which visualizations are affected?
  • Which users are affected?

Research questions

• What are the requirements for a visual impact analysis and simulation?
• How to determine and visualize dependencies between data, metrics, visualizations, and users?
• What are possible types of data and changes in the context of enterprise metrics and visualizations?
• How to visualize changes (especially) of complex linked data (e.g. networks, trees, etc.)?
• How to design an interactive user interface for the simulation of changes?
• How does a prototypical implementation look like?

Approach

1. Elucidate requirements
2. Literature review on visual impact analysis
3. Solution design 
4. Solution implementation
5. Case evalutation (with industry partner)

Expected outcome

• Running prototypical web application for the visual analysis and simulation of data changes to enterprise metrics