I am doing some correlation study between some tool interruption and the number of production moves. For example in one day I have a 5 interruption versus a total production moves of 1000. What I observed is that when I do a correlation analysis I was not able to get a good correlation between the interruption and moves after analyzing a total of 9 month of daily data. But I can see that sometimes that the trend actually follow the number of production moves. How do I perform a better data segregration in order to see how many of instances that the interruption is actually correlated by production moves. I have tried the cluster analysis, but it is not able to indicate a cluster that is highly correlated.