These charts were created using Microsoft Excel, based off of US census data.
Recently I observed a poster detailing a study of Visualizing Multi-Agent Collaboration for Classification of Information. The data presented showed that over time, collaboration among agents actually decreased in successful ventures because the agents grew to work more efficiently on a task, and therefore fewer were needed to successfully complete each one. The chart also showed many unsuccessfully completed tasks kept a larger number of agents working to collaborate, and presumably, not learning or adapting towards efficiency. “Wow,” you might say. “What kind of a chart could tell me all of that information at once?” Well, I’ll tell you–a bar chart with two X-axes, with yellow bars showing tasks uncompleted, black bars showing tasks successfully completed, the Y-axis measuring the number (1-25) of agents involved in each task, and a white line graph overlaid measuring the average collaboration to success ratio.
It was useful to be able to see the successful and unsuccessful collaborations on the same chart; it made it very clear that the change in collaborators and success rate showed an increase in efficiency among some agents, but no change in other agents. The colors chosen for the chart were stark enough to be clearly distinguished, and although the bars themselves were very narrow, the chart was large enough that viewers could still read all three sets of data clearly. What might have been helpful is more description of the tasks that these agents were completing successfully or unsuccessfully; I can only imagine that this information was part of a presentation that accompanied this chart and did not make it onto the poster.