Testing for cyclicality in data

#1
Hi

I have a set of data - quarterly figures of a company's financial statements - and I would like to see whether the figures are cyclical e.g. 4th quarter always the highest. Which statistical test should I apply and whether such a function is available in Excel?

thanks
 
#2
One possibility could be to use code the first quarter (Q1) as reference category and the rest quarters as dummy variables and test if the effect for Q4 (fourth quarter) is larger than the other ones.

(That is not what you explicitly asked for. What did you ask for? There are other difficulties with your data that you need to discuss.)

Thanks for not having "help" or "please" in your title. Such titles are very bad descriptors and are seldom read and answered. If you want to have an answer make an informative title!
 

Miner

TS Contributor
#3
I believe that you may be confusing cyclicity with seasonality. Seasonality is a repeating event such as you describe. It can happen at an hourly, daily, weekly, monthly quarterly, etc. level. Cyclicity is by nature unpredictable except for a very short forecast (i.e., less than 3 months). There are ups and downs but the duration and frequency change.

Seasonality may be analyzed using decomposition or by cross-correlation of the time series with itself. If you are using ARIMA, seasonal patterns will often show up in the ACF and PACF as a suspension bridge pattern, the length of which is the seasonal period.

Regarding Excel, the only possibility that might work is the Fourier Analysis and picking the seasonal period from the Periodogram. Theoretically, it should work and I have found some articles using a search engine. I have used Fourier Analysis on sound and vibration, but never on a time series.