If you use Matlab:

http://www.mathworks.com/matlabcentral/fileexchange/20329-cosinor-analysis/content/html/cosinor.html

You can check Dr.'s Roelof A. Hut homepage:

http://webpage2.woelmuis.nl/index.htm

There you can find the Cirwave software and other programs for rhythms analysis.

There is also the ActoJ module for Image J software that can help you with time series analysisi for biological rhythms studies.

I use cosinor to calculate amplitud, acrophase and mesor.

For circadian rhythms I use a software from this page.

http://www.chronopharmacology.de/software.htm

"Chronos Fit" software its free to use. you need registration. It gIves you if you have or not a circadian rhythms.

i think it is OK using Matlab software just for analysing circadian period.

sorry,I don't know whether the matlab is free or not in Greece. 

http://www.mathworks.com/matlabcentral/fileexchange/20329-cosinor-analysis/content/html/cosinor.html

In my experience it is simplest to perform cosinor analysis in R. I has packages called "season" and "cosinor" both of which can easily be installed. You dataset should have a numerical response variable and a vector for time, which are the only two quantities you need for perform this analysis.

Another program to use is the       GraphPad Prism 5       with "XY analisys",   "Nonlinear Regression (curve fit)".

Using this Equation: Y=Baseline+Amplitude*cos(Frecuency*X+Phaseshift).

Rules for initial values

Baseline: 1      RULE: *YMID

Amplitude: 0.5     RULE: *(YMAX-YMIN)

Frequency: 0.2618      RULE: (Initial value, to be fit)

Phaseshift: 1.0      RULE: *(Valueof X at YMAX)

Defaults contrainsts

Frecuency Constrains equal to: 0.2618

Baseline, Frecuency and Phaseshft : NO CONSTRAINTS

Any questions?  [email protected]

You obtain the next results. 

"Best-fit values"

" Baseline"

" Amplitude"

" Frecuency"

" Phaseshift"

"Std. Error"

" Baseline"

" Amplitude"

" Phaseshift"

"95% Confidence Intervals"

" Baseline"

" Amplitude"

" Phaseshift"

"Goodness of Fit"

" Degrees of Freedom"

" R square"

" Absolute Sum of Squares"

" Sy.x"

Constraints

" Frecuency"

"Number of points"

" Analyzed"

You can use R package 'cosinor' for the analyses:

https://cran.r-project.org/web/packages/cosinor/cosinor.pdf

I hope it helps! 

For an analysis of a rhythm (annual or daily) the first step is to demonstrate that this rhythm exists and is stable. It is not the topic of cosinor method. The anova with repeated data is classicaly used. You have to compare the level observed for example during night or during light phase during more than 2 days. After, you can use cosinor method. A non-linear regression on y=a+b sin(wt+f) can be also use to study this rhythm it more interesting for a lot of point but it is more difficult to use.

you can try pyActigraphy (python)

https://ghammad.github.io/pyActigraphy/pyActigraphy-Cosinor.html