Click the Lomb-Scargle tab.
The Lomb-Scargle periodogram essentially fits a sinusoidal model to each frequency, where the tested set of frequencies depends on the sampling interval and the length of the time series. To improve the resolution, more frequencies can be used, set by the oversampling factor. Too high of an oversampling factor will have diminishing returns; a value 4-8 is considered reasonable.
The Lomb-Scargle periodogram calculates a p-value for the dominant period. The summary table includes the percent of samples with p<0.01, indicating clearly discernible rhythmicity. If the noise standard deviation is set to be very large compared to the amplitude of the time series, then some time series might not pass this rhythmicity test (signal too buried in the noise). The p-value listed in the table and shown as a horizontal orange line on the periodogram is that given by the theoretical formula.
A randomization procedure for computing a p-value is also given (shown as a vertical orange line in the histogram above), which might be desirable if the time series don’t meet the assumptions of the theoretical formula. This procedure tends to be very time-consuming, so is only done for the representative time series; the summary results use the theoretically computed p-value.
Once you have used the app to learn more about the Lomb-Scargle periodogram, answer these questions to test your knowledge.