Rule-based Evaluations for Mobile Data Collection Applications

Paper-based data collection is often directly associated with high cost and high delays. Collected data is often shipped to domain experts in order to be analyzed. This, in turn, leads to waiting times before results are available. However, in many clinical or psychological cases, e.g., with the purpose to detect potential risk factors, real time analysis would be a major improvement for all stakeholders. To achieve the latter an electronic data collection approach is needed instead of a paper-based one. A framework was developed to allow domain experts to create mobile data collection applications by themselves. This framework is shortly introduced in the beginning of this thesis. The framework makes the creation of questionnaires much more efficient. Domain experts are able to create mobile applications without needing IT expertise. Furthermore the framework allows the definition of so-called rules. Rules provide the possibility to analyze the answers of a questionnaire. To increase the efficiency of this framework further, this thesis will create a tool providing the functionality to automatically evaluate these rules with the result of the questionnaire in order to provide real time analysis and feedback. This tool is meant to be integrated into the existing framework and executed on a participant’s mobile device or browser. Since rules may contain malicious code, they have to be executed in some sort of a secured context to provide safe evaluation. After comparing several different approaches, the tool was implemented with the use of the best fitting solution. In order to evaluate rules, necessary information is extracted from the questionnaire result and model and a sandbox with variables and functions build. Then a single rule is evaluated without access to any data outside its own context. Altogether this thesis implements a tool as an addition to the framework to provide safe real time analysis.