Animated Visualization of Frequent Itemsets as Hypergraphs
If you simply want to view the resulting visualization, you can check out the results here.
If you want to play around with the graph's implementation and build the dist-file you need to have
Node.js and gulp.js installed. The build-task is defined as the
default task, so you can simply run gulp
from your shell of choice in the project's directory and you're good to go.
The version number for the dist-file is taken from package.json
.
This Proof Of Concept was developed as part of a student research paper at HSR in 2015.
The visualization of Big Data is a non-trivial affair. As the type of data and information about it are not known in advance, it is complex to find a suitable and optimal visualization technique. In Prof. Dr. Eduard Glatz's paper Visualizing big network traffic data using frequent pattern mining and hypergraphs it was shown, that through Data Mining calculated Frequent Itemsets can efficiently be displayed as a hypergraph with different vertices. Although hypergraphs might no be the most suitable visualization technique in all cases, they provide a generic and flexible solution.
In this student research project various layout algorithms were developed and tested in a experimental manner, which are suited for changes over time. Thereby are some of the most important factors overlapping of nodes and links and similar problems. These problems were also considered and attended to in the student research project.
The result was a layout algorithm which displays a structured hypergraph in a browser and represents the chronological dimension with animated toggling of visibility. Due to the use of web technologies a user can interact with the hypergraph and its animation, e.g. controlling the time interval. Furthermore related concepts where examined and documented, such as various layout algorithms and alternative, interactive visualization techniques.