Data Visualization for Discrete Event Models

This thesis combines web technologies, data visualization and discrete event simulation.  The result will be a generic, user-friendly way of visualizing data from a running model.  

The ABS language, developed by the ASR group, enables users to model server load and bandwidth consumption for different workloads, movement of vessels off the coast of Norway, behaviors of celebrants during New Year’s Eve, and anything else interesting that can be expressed as discrete events happening at a specified time.

ABS models produce data, usually in the form of time series.  Each ABS model contains a small web server that makes this data accessible via JSON.  Currently, visualizing this data means writing a bespoke page of HTML and Javascript that reads JSON data from the model and displays it in a browser window, usually using the D3 Javascript library.

The result of this thesis will be a general visualization frontend for ABS.  The user should be able to choose multiple data points and time series from a menu, choose how to visualize them (line vs bar chart, colors, position of graph, ...), and save these choices for the next time the model is run.

Work on this thesis includes:

  • Selecting frontend libraries and visualization technology
  • Implementing the visualization frontend
  • Documenting best practices for model authors on how to get best visualization results
  • Creating some sample models as teaching aids
     
Emneord: data visualization, JavaScript, JSON, web applications
Publisert 30. sep. 2020 14:42 - Sist endret 30. sep. 2020 14:42

Omfang (studiepoeng)

60