Ömer Bayram:

Visualization of Linked Data Entities

Requirements

  • Very good knowledge of JavaScript
  • Knowledge of creating visualisations with JavaScript, e.g. with D3
  • Knowledge of Linked Data and RDF is benefitial, but not required
Academic Advisor
Discipline
Web Development, User Interface Design, Interaction Design, Usability, Human Computer Interaction, Linked Data
Degree
Master of Science (M.Sc.)

Contents

Context:

Linked Data is a general principle for describing and publishing information about entities of an application domain. The description of Linked Data entities is implemented in a text-based machine-readable format that follows the Resource Description Framework (RDF).

The signature of a Linked Data entity is comprised of its attributes (data properties) and its typed relationships (object properties) to other entities.

Objectives:

In this thesis, you will develop a visualisation of the Linked Data entities in two or three dimensions based on their signature. The shape of the entity corresponds to the signature of the entity.

The resulting entities could for example look like Tetris bricks. This would allow to visually classify, cluster, sort, recombine and otherwise interact with the entities.

One could imagine forming composite structures with these bricks, each having a unique signature. For example, a “family” could consist of two “parent” entities and at least one “child” entity.

Related Work:

Linked Data entities have been visualised in a number of applications.

An example is the DBPedia atlas [1], which visualises the entities of DBPedia. However, in this visualisation, as in many others, the individual items all have the same shape.

Another example is the three-dimensional visualisation provided by HDT-it [2], which visualises an RDF graph in three dimensions.

A different example are cellular automata [3]. The items in such simulations form a certain shape. Additionally, they also autonomously move and interact with other shapes, forming new shapes and entities that have been attributed a specific semantic meaning (e.g. a “glider”).

Procedure:

  • Selection of a dataset
  • Visualisation design
  • Interaction design
  • Implementation as a prototype
  • Rigorous qualitative and quantitative evaluation of the prototype with subsequent analysis
  • Comparison of results with other applications
  • Testing theses in comparison with recorded metrics