02-01-2011 04:48 AM
Jeff, as you already have the data in Excel, could you do some experiments?
First of all, I think we should apply some scaling to both axes so the max is 100%.
Second, I think it's better to use the integral (sum) of the kudos. If we then get 80% Kudos at 20% 'groupies' we could assume a pareto statistic and try to fit the data to extract the parameters.
Felix
02-01-2011 01:23 PM
@f. Schubert wrote:
Jeff, as you already have the data in Excel, could you do some experiments?
If I may make a small recommendation to Jeff, there is a fabulous piece of software that you could use to make this analysis easier for both you and the rest of us, it is called LabVIEW. Perhaps you have heard of it, not as prevalent as Excel, but very nice once you have learned how to use it. As a bonus, the graphs are slightly less unsightly.
02-01-2011 04:38 PM
HEY-
CHEAP SHOT!
OK - I was looking for broad brushstrokes hand loading some sample data since I'm too busy working to write a data mining app. (and did not save the raw data either since it was not fully audited)
BESIDES- I thought to leave the detailed statistical treatment to the pros.