Not sure what you are doing exactly, but here is a little booklet that explains very simple interpolation and extrapolation (on graphs) using regression calculations in excel.

Not sure, if the structure of the formula remains the same you could fill in the formula with data and then make a graph out of the formula instead of data? If it's random data I wouldn't know.

Given a set of related data that are related in some way (linear or quadratic...in this case it's quadratic) finding an unknown value (say on the x you know 1,2,3..I'm looking for 4) EDIT: Kreij gave me the right stuff...thanks you guys.

This may sound daft as I'm not sure about the accuracy of the excel trendlines, but what you could try is right clicking on the line of the data you want to extrapolate from, making an appropriate trendline (I dont know what sort of line ur dealing with, so pick from the options whichever type of trendline most closely fits your data), then in the options on the same menu as from where you choose ur trendline, in the box marked "Forward" or "Backward", enter how many data points forward or backward you wish to extrapolate. It <i>should</i> be reasonable accurate, but I havnt come across any software yet which can extrapolate decently, so you may be better off working it manually, then include the extrapolated data in your graph data range. Apologies for the poorly worded explanation and the utter lack of full stops, I couldnt think of how to word this better

Yeahh that's what I ended up doing. Also I figured out that if you need an extrapolated piece of data at a certain point you could get excel to give you an equation for the trendline and use that.

Good stuff mate. Glad it went all right. What were you extrapolating btw (curious from the perspective of a physics student)?