Extrapolation of graph for Trend

[RRMummy]

Hi techies,


I need help here. I would like to predict the trend of something for the next 2 - 3 years.

1. Besides doing a manual calculation, is there a function in excel which will allow me to do that?

2. Will this then give me a graph which will allow me to use --- line format for the extrapolation part.

Thanks in Advance. :celebrate:

[duriz]

Hi RRMummy, I hope this is useful


1. Start Excel.

2. Enter Data for graph: When you have all your data entered, select both columns of data and click on the small bar-graph icon at the top of the window. This is the Chart Wizard. Click Next to see a sample of your graph. Click Next again to continue. When you have completed the details for your legend, click Finish to see a completed graph.

3. Extrapolating Data:
This is a way to extend your trendline so that additional points can be predicted or an x- or y-intercept can be found. To extend your trendline, select your trendline and from the Format menu, choose Selected Trendline. Click on the options tab and then use the Forecast option to extend your line backwards or forwards. If need be, check the box that says “Display equation on chart” to show the slope-intercept form of a linear equation. (This will show you the exact y-intercept.)

[RRMummy]

Thanks for the quick reply duriz. I'll go try it now.


Thanks :celebrate:

[duriz]

LKVM:

Wow Duriz apart from cooking tips u give nice excel tips too :salute:

No lah.
I work on Excel at work everyday, for the last 5 years. Churning out report after report.
It's been a love-hate relationship

But I think faster than I type. I hope it all made sense to RRMummy :oops:

[limlim]

What if the trend is non-linear :evil:

[duriz]

limlim:

What if the trend is non-linear :evil:

You can add a nonlinear trend line to the chart by selecting your chart and going to \"add trendline\" on the chart menu. You can select a non-linear sort such as a polynomial.

If you select display equation on chart under the options tab you can see the formula for your line.

The various options are:

Logarithmic: This is a linear-log relationship.
Exponential: This is a log-linear relationship
Power: This is a log-log relationship.
Polynomial: This is for example a quadratic relationship if the poynomial is of order 2.
Moving average:This is generally only useful for time series data.

[limlim]

The problem is, you need to know the "nature" of the data (e.g. polynomial) beforehand before deciding which trend line to use… ok, juz joking regarding the forecasting for non-linear trends… I guess non linear curves are not so straight forward… it’s not easy to judge the "type" by inspecting a few data points or even a few hundred data points… (^.^)!


By the way, I doubt MA can extrapolate trend lines or do "forecast" leh…