Forecast forward using observed data

[tinkerbelle]

Hi, please see attached spreadsheet.

The deseasonalised month and day are combined to create the combined view of that day.

I have 731 days historical data.

I would like to be able to forecast forward (eg: what will happen on day 732 or 780 for example).

Can you explain what I need to do? I have been looking up LOGEST, Forecast, Trend etc but not sure where to start!!

Many thanks for your help.

[MrShorty]

I often find that the best first place to start on something like this is not with Excel.

1) Because I believe it is important to understand the math behind what you are doing before you can program it into Excel (or other programming language), I would suggest reviewing regressions. This http://www.purplemath.com/modules/scattreg.htm is a nice tutorial for high school algebra students about scatter plots and regressions.
2) One of the first steps in regression is to decide what kind of equation will provide the best fit. I will usually start by plotting the data in a scatter plot and looking at it. The chart has the ability to add trendlines, which can be a useful tool to quickly experiment with different types of equations. http://office.microsoft.com/en-us/ex...158.aspx?CTT=1 Once you have settled on a suitable trendline, you can then have the equation displayed, manually enter the coefficients for the equation into the spreadsheet and calculate the desired data from the equation. For those unfamiliar with the spreadsheet functions, this can be easier than trying to learn them.
3) If this is a one time task that you will never or rarely do again, that might be enough. If this is something you need to become proficient in, I would suggest learning how to use the LINEST() function to perform regressions in the spreadsheet.

[tinkerbelle]

Thanks very much for the reply, I will have a look at the links.