Do you have an underlying model for the data? Or do you not have any clues as to the type of function the data is supposed to represent?
Ah. That doesn't surprise me in the least. And yes, that is a semilog plot. Probably the easiest thing to do is to fit a straight line to the semilog plot of your data. The idea is this: assume your function (on the original plot) is of the form
Here it is essential that both and
Now, plotting in a semilog plot is the same as taking the logarithm of both sides of this equation. The result is the following:
Your semilog plot has essentially plotted versus Therefore, you can find the slope and intercept of a straight-line-fit. The slope will be , and the intercept
Make sense?