How to add a trendline to a log graph when x=0?

[jz9]

Hello,

I tried to plot my data (attached) and it looks like a log graph. I wanted to add a trendline, but it is not possible since the data starts at x = 0. At x = 0, it has a value (y-axis) and this is where the percentage of a substrate is 100%.

How may I plot the graph if I want to have a trendline equation? Should I start plotting at the second reading i.e. after x =0? If so, how I may extend the trendline so that the y-intercept is the y-value of x = 0?

Thanks for your response, I really appreciate it.

[MrShorty]

I don't know what kind of regression equation/trendline you should be using here. If you need the regression equation to be valid at x=0, then you cannot use something with a log(x) term in it. IMO, the question seems to boil down to choosing a suitable regression/trendline quation, which is not a question that Excel can answer for us. You as the programmer/analyst need to know enough about the process you are modeling to choose an appropriate model before performing the regression.

My own quick look suggests that ln(x)=my+b could be a decent fit to the data (except for the x=0 point), but there is no way to extrapolate that equation out to x=0.

A hyperbolic regression equation (y=A+B/(x+C)) might work, too, but that is a non-linear equation, so you would not be able to use the chart trendline for that one.

But, at this point, I am just guessing at possible regression equations. What regression equation do people typically use for this analysis? If they use something involving ln(x), how do they typically handle the x=0 points?

[jz9]

OK, understood. Thank you very much for your response, I appreciate it!