Periodograms are essential for definitely showing periodicity in a quantifiable way. Periodograms often make use of algorithms created for spectral analysis, to decompose a signal into its component waves of varying frequencies. This has been adopted to behavioural data, in that it can find a base rhythm over several days from what is usually unclean data.

Ethoscopy has 5 types of periodograms built into its behavpy_periodogram class, which are 'chi squared' (the most commonly used), 'lomb scargle', fourier, and 'welch' (all based of of the Fast Fourier Transformation (FFT) algorithm) and 'wavelet' (using FFT but maintaining the time dimension).

Try them all out on your data and see which works best for you.