Plot trend line from data imported from excel

Hi, I have used matlab for some time now, for importing data from excel and plotting it in matlab because the graphs look much better. This is the first time I have to do a trend line of the data. What is the way to do it because in insert there is no trend line option. In excel this is much simpler.

Do you have any idea how to do it?

I want to find trend line for the orange line separate from the blue line as well as a combined trend from both data sets.

I have attached a photo you can see the data there.

Best Regards,

Tsvetan

[Star Strider]

Originally posted as https://www.mathworks.com/matlabcentral/answers/344206-not-a-qustion-anymore

Best to have clarified and added to the original post rather than posting a duplicate Question.

Answer 1

[dpb]

It's not much more than Excel altho a "trendline" is generally considered linear and certainly the orange curve isn't going to be well represented by such. But, basics are the same other than it takes more than the two endpoints to draw a smooth curve of higher degree than a straight line.

I'm going to use shorter variable names, though... :)

bUK=polyfit(year,UKdata,1);  % trendline coefficients UK
bWW=polyfit(year,WWdata,1);  % ditto worldwide...
hold on
yrs=[year(1) year(end)];     % first/last year in dataset
hLnUK=plot(yrs,polyval(bUK,yrs));  % UK trendline
hLnWW=plot(yrs,polyval(bWW,yrs));  % WW trendline

Set color, line style, etc., as desired...

If it's desired to annotate the lines, that's easy enough, too...

hTxtUK=text(year(ix),dataUK(ix),num2str('UK = %.1f*Year+%.1f',bUK);

where ix is a year index chosen to put the text at a convenient location; hard to choose when the plot is occluded by other windows. One can see there's a big spike at the RH end for WW; one presumes something similar for UK as well but not as pronounced.

"Salt to suit..." but should get you started...see the documentation of course for more details on all the above...

Answer 2

[Star Strider]

From my answer to your previous Question:

If you just want to plot a linear least-squares fit to your data, see if the lsline (link) function will do what you want.

The polyfit and polyval functions can create polynomial fits to your data, and will return parameters for the fit.

Answer 3

[Tsvetan Donov]

Thanks a lot! I will try it and see if I can manage to make it work. Now very good with code..

If I want to plot multiple data do I put

plot(x,y,x,y,x,y)

In which case plot(Year,bUK,Year,bWW,Year,UKdata,Year,WWdata)?

Well I can later annotate them in the user-friendly interface :)