BSV

A visualization tool for NetLogo BehaviorSpace runs.

Goals

• Import arbitrary CSV in intelligently, allowing easy access from NetLogo code.
• Visualize data as points with attributes (xcor, ycor, color, label, size) set to arbitray reporters based on data.
• Intelligent defaults for attribute scaling.
• Easily transition between different visualizations.
• Group data based on arbitrary keys.
  • Example use case: Be able to visualize mean values for repeated runs.
• Ability to visualize a single row as multiple points, each point using different columns.
  • Example use case: Suppose you have data from a single run tracking three populations. You want to be able to plot each population over time simultaneously.
• Mix grouping and splitting
  • Example use case: Have multiple runs of three populations. You want to be able to plot the means of each population across the runs simultaneously.
• Build-in clustering algorithms.
  • Example use case: Color clusters in scatter plot.
  • Example use case: Use clusters as groups, so you could get means within clusters.
• Be able to handle 2 uniform inputs with one output. A heatmap would be one way to do this.
• Visualize changes over time.
  • One simple thing to help this is just add a connect-to reporter, which woud create a link from a point to another point if the reporter is true for that point.
  • A more sophisticated way would be "sliding window" points.

Ultimate test case: Be able to visualize model state space, averaged across runs. Be able to visualize the vector field of the state space would be pretty awesome too.