Written by craig kemp on Wednesday August 25th 2010 in What We're Thinking
‘Words divide, pictures unite’
Ever get the feeling that you are suffering from a data overload?
Ok, let me set the scene, I recently built a model, ran it, modified it, ran it again, tweaked it a bit, ran…you get the idea. So anyway, I am now sitting here staring at results from 10 key runs, with 3 KPIs each. That’s 30 individual pieces of data staring back at me. Which run and model set-up is optimum? Well, where do I start? Run 7 had excellent Throughput, but the Resources were very underutilized, run 3 had good Throughput and used the Resources well too, but then so did run 9…
I decided to turn my attention to visualization. I exported my data to a visualization package and created a scatter plot to compare each run, the x-axis, the y-axis and the size of the plot on the graph all corresponding to one of my KPIs. This allowed me to have a visual display of all my results without getting overwhelmed.
With data visualization being such a ‘hot topic’ these days, and even being reported on mainstream news shows earlier this month, it got me thinking about how data visualization could play a role in the development of discrete event simulation data, in particular, how we can visualize our results to cope with the masses of data that are produced at the end of each run, and then allow visualizations to provide us with an even deeper understanding of our models.
How do you visualize your results? Bar charts, Line graphs, maybe Radar graphs? Or maybe none at all!
It would be great to hear about your visual solutions, why not send them in?