Using excel, how can I make a graph that curves off as it moves towards an Asymptote?

[webnewland]

I have bunch of data points, and looking to have the graph curve off towards an asymptote rather than having it make a line of best fit, how can I do it?
Looks like this
Link

Thanks in advance

 

[darktubbly]

This is a cheesy workaround, but add additional data points that are on the asymptote to your chart's source data. After you create the graph, you can add a trendline, and then right-click the extra points and set them to have no marker visible.

 

[raptor13]

Instead of doing a linear trendline, do an exponential or polynomial trendline and set the power as high as it will let you set it. Yes it's only an approximation of your data but I'll be surprised if a 6th order function can't very nearly describe your data set. That is especially true if all you want to show is the shape of the curve approaching its limit.

 

[webnewland]

Originally posted by: darktubbly
This is a cheesy workaround, but add additional data points that are on the asymptote to your chart's source data. After you create the graph, you can add a trendline, and then right-click the extra points and set them to have no marker visible.

The thing is I only have data points, I can see that points are leveling off at the top, but I don't want to arbitrarily assign an asymptote, There is no way to tell excel to draw an estimated curve similar to line of best fit?

 

[raptor13]

More details:

Select the data set in the chart window. Then do Chart/Add Trendline. Select the power, exponential, logarithmic, or polynomial fit line. Crank the function order up. Enjoy!

 

[webnewland]

Originally posted by: raptor13
Instead of doing a linear trendline, do an exponential or polynomial trendline and set the power as high as it will let you set it. Yes it's only an approximation of your data but I'll be surprised if a 6th order function can't very nearly describe your data set. That is especially true if all you want to show is the shape of the curve approaching its limit.

I tried to do an exponential trendline, but it looks like this
link, it leveled off too early, the polynomial trendline has a too many bumps.

 

[raptor13]

If you only have seven data points, good luck trying to make a good function out of that. Judging by your axis labels, I'm sure you're doing this for school so the theory behind whatever you've done is well known. However, your data sure doesn't look like it's approaching a limit! If you're absolutely set on making your known result fit your data (that's some good science! ) then your best bet is as follows:


View/Toolbars/Drawing


You can take it from there.

 

[webnewland]

Originally posted by: raptor13
If you only have seven data points, good luck trying to make a good function out of that. Judging by your axis labels, I'm sure you're doing this for school so the theory behind whatever you've done is well known. However, your data sure doesn't look like it's approaching a limit! If you're absolutely set on making your known result fit your data (that's some good science! ) then your best bet is as follows:


View/Toolbars/Drawing


You can take it from there.

Heh ya the lack of data points is a problem. I guess I'll add in some fake markers and make it draw an asymptote lol.

 

[raptor13]

Originally posted by: webnewland

Originally posted by: raptor13 If you only have seven data points, good luck trying to make a good function out of that. Judging by your axis labels, I'm sure you're doing this for school so the theory behind whatever you've done is well known. However, your data sure doesn't look like it's approaching a limit! If you're absolutely set on making your known result fit your data (that's some good science! ) then your best bet is as follows: View/Toolbars/Drawing You can take it from there.

Heh ya the lack of data points is a problem. I guess I'll add in some fake markers and make it draw an asymptote lol.

Another thing you might try is just doing the best fit line you can with the data you have. You can then discuss how your gathered data does not support known theory. You can then go on and on about deficiencies in the experiment, need for further study, and all sorts of stuff you don't actually care about.

My professors always ate that crap right up!

 

[silverpig]

That actually looks like a decent fit for that data.