The bump you see there is 12 hours. It doesn't give a whole lot of information about how to make websites better but it does show that internet traffic is pretty bimodal. Visiting once before work and once after work. Which tricks the data into thinking its on a 12 hour cycle.
That or I don't know how to read the graph and it's 24 hours.
I used this: https://www.npmjs.com/package/fourier-transform
The fft comes out with half the length of the source data. The length of sampling is 3 months. So I figure 3 months over 182 (the index of that point) is 12 hours.
I was suprised I didn't see anything at a week.
https://jssocial.pw/ppkey/fget/x0x7/upload/fft.txt ; Compatible as CSV
The bump you see there is 12 hours. It doesn't give a whole lot of information about how to make websites better but it does show that internet traffic is pretty bimodal. Visiting once before work and once after work. Which tricks the data into thinking its on a 12 hour cycle.
That or I don't know how to read the graph and it's 24 hours.
I used this: https://www.npmjs.com/package/fourier-transform
The fft comes out with half the length of the source data. The length of sampling is 3 months. So I figure 3 months over 182 (the index of that point) is 12 hours.
I was suprised I didn't see anything at a week.
https://jssocial.pw/ppkey/fget/x0x7/upload/fft.txt ; Compatible as CSV
(post is archived)