Periodic patterns in Web log time series

Recently, I have analysed trends about visitors on this blog. I have made two observations. First, there is about 500 to 1000 visitors per day. For this, I want to thank you all for reading and commenting on the blog.  Second, if we look carefully at the number of visitors per day, it becomes a time series, and we can clearly see some patterns that is repeating itself every week. Below is a picture of this time series for January 2018.


periodic visitor accesses

As you can see, there is a clear pattern every week. Toward the beginning of the week on Monday and Tuesday, the number of visitor increases, while around Friday it starts to decrease. Finally, on Saturday and Sunday, there is a considerable decrease, and then it increases again on Monday. This pattern is repeating itself every week. We can see it visually, but such patterns could be detected using time series analysis techniques such as an autocorrelation plot. Besides, it would be easy to predict this time series using time series forecasting models.

We can also see a relationship with the concept of  periodic patterns that I have previously discussed in this blog. A periodic pattern is pattern that is always repeating itself over time.  That is all for today. I just wanted to shared this interesting finding.

==
Philippe Fournier-Viger is a professor, data mining researcher and the founder of the SPMF data mining software, which includes more than 100 algorithms for pattern mining.

(Visited 34 times, 1 visits today)
This entry was posted in Data science, Time series, Web and tagged , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *