Trying to evaluate high variance in large dataset

44 Views Asked by At

So I have an large active dataset, recording speed tests of connections. I am looking for the best approach to identify possible problem connections.

To do this I started by looking at averages etc. This works in a very simple scenario and is able to highlight problems in the moment, but I want to track high swings over time that indicate possible hardware issues etc.

I have tried using standard deviation, but I admit to being new to this.

Today I use the following:

  1. Overall average speed
  2. average of recent tests against overall average
  3. Standard Deviation rating.

What I am looking to catch:

The problem connection are typically represented by peaks and valleys in the data where speeds are cutting in and out.

Is there a recommended approach?

Thank you for your attention.