In seiner Funktionalität auf die Lehre in gestalterischen Studiengängen zugeschnitten... Schnittstelle für die moderne Lehre
In seiner Funktionalität auf die Lehre in gestalterischen Studiengängen zugeschnitten... Schnittstelle für die moderne Lehre
Semi-automated rule based data annotation is a frequent pattern that emerges in all kinds of user interfaces. The example that is probably the most wide-spread is email filtering, where a user sorts incoming email into different folders on their system. Typically this is a process that necessarily contains manual decisions done by the user on a case by case basis, while also some email can be categorised automatically by applying rules. For the user to be able to define helpful filtering rules, the interface needs to be expressive enough to allow for different use-cases, while being restricted enough to not result in filtering rules that are impossible to apply.
In this thesis I investigate this data annotation workflow and how the user interface for designing the filtering rules can be built. My software implementation is able to sort bookkeeping transactions into different account names, as it is a typical workflow for the bookkeeping department in small associations and corporations.
Rule based data annotation is an emergent pattern in UI design and its implementations can be compared in regards to their effectiveness to communicate the relationship between data and rules, learning curve and domain fit. In my application I found the following considerations to be most important.
I think that rule-based data annotation can be a helpful interface pattern to reason about. It can be applied in any environment where users need to annotate a stream of data items, where the data can be meaningfully categorised by rules.