I have a worksheet with hourly data for one year in it (8760 rows, one row per hour). The X column is the Date (i.e.1/1 in date format, 2444605 in numeric), Y column is the hour (0 through 23 for each day) and the Z column is a temperature reading. I would like to create a 3D surface plot of this data. My goal is a graph that just 'connects the dots' of the 3D xyz scatter plot with plane surfaces. I have a few questions:

I assume I need to create a matrix to do this, is this correct?

Is there anywhere in the manuals that explains the error messages from the convert to matrix routine on the edit-convert to matrix-random xyz? The weighted average option gives "Gridding error = -1" and the quick correlation gives "Error of inversing matrix."

Using correlation (and waiting -- maybe this is the excuse I need to squeeze the boss for a faster computer , I can create a matrix and then a surface plot. What would be the best matrix size (considering my goal above) to use to show the 8760 data points? Is the thinking 24 rows (for 24 hours) and 365 columns (days)on the right track? I read a post regarding surface plots at http://www.microcal.com/bulletin_board/Forum1/HTML/000506.html, but where I am confused is how to optimize the matrix to take advantage of the fact that there is only one random variable (z, ranging from 0 to 90)?

Now a couple quick (I hope!) graph formatting questions. Once I get the surface graph, the date axis is shifted. Rather than Jan 1981 - Jan 1982, it is Aug 1980 - Feb 1982. Is the data 'streched' by the conversion to a matrix, or is it just being extrapolated out (so I should just limit the scale and ignore the bit chopped off)? Any ideas?

And finally, the surface plot displays gridlines that follow the surface. In Plot Details I can turn the grid on or off, but how can I increase the number of gridlines to show a tigher mesh fitted to the surface? They don't seem tied to the axis settings.

Thanks for all the help, in the past this forum has proved to be a lifesaver!

Your data sounds like it is regular XYZ data. There are a certain number of ordered X values for each Y value, and the same number of ordered Y values for each X value. To convert regular XYZ worksheet data to a matrix, select the Z column and then select Edit:Convert to Matrix:Regular XYZ.

You do not need to perform a random conversion because, as you say, the only random variable is Z.

When you perform a regular conversion, the matrix size is automatically set up.

There is an explanation of the error message that you were getting in our technical support Knowledge Base, Convert XYZ to Matrix causes Error.

Your other questions regarding the formatting of the graph are irrelevant once you convert your data as regular data.

If you need further assistance, please email me directly at tech@microcal.com.