Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 0 down vote favorite

I am trying to track the performance of a set of homogeneous subsystems vs the system as a whole. The performance metric is measured by a mean value based on a set of tasks performed by a the subsystem, for example, the mean wait time before scheduling task.

Each subsystem may be responsible for varying numbers of tasks. I can calculate the mean wait time for each subsystem and for the system as a whole.

My question is how to measure and visualize how each subsystem is performing compared to the whole system average (or maybe there's a better aggregate measure for the whole system?)

Obviously being below the mean shows the subsystem is performing relatively better and above the mean vice versa, and the distance from the whole system mean represents that, but I struggle to get real meaning or provide a good visualization from that.

Is there a better way? Or if I'm on the right track, how do I assign meaning / visualize the metric?

share|improve this question
2  
It will help if you can tell us what you are trying to accomplish ultimately beyond the specific visualization difficulties you have at present. – user28 Nov 24 '10 at 14:23
When I look at the system performance metric I want to understand how each subsystem is contributing towards it so I can get an idea of how significant the difference is. Ultimately I'd want to work out the tolerance and raise a red flag when subsystems perform worse by more than that tolerance. In a healthy system, I'd expect the the subsystems to be delivering approximately the same and therefore the subsystem values should be spread relatively closely around the system value. – fd01 Nov 24 '10 at 15:15
In summary, I want to recognise when a subsystem is performing badly relative to its peers. – fd01 Nov 24 '10 at 15:20

1 Answer

up vote 2 down vote

It sounds like you should lookup the topic of statistical process control. Its origins lie in manufacturing where deviations from expected process output are used to raise flags to ensure that quality of manufactured items is as per pre-established norms. In particular, see the wikipedia description for control charts which should get you started on the visualization aspect.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.