Data analytics and visualization of Oura biosensor data

Hi, new to this forum. I recently downloaded my biosensor data from Oura after 800 days of wearing the ring. Then did some multivariate analyses and graphing. Some interesting patterns.
I’d love to see if other people also see similar patterns in their data, and if you have suggestions for better algorithms. I used a software called JMP, but I’d also love to see these types on analyses and graphing done using an open-source like R. (below is the link to the blog post)

Finding patterns in my days: a visual exploration of my Oura data

7 Likes

This is amazing @BeniM! Thanks for sharing.

  • How did you decide which day to assign the sleep data? Let’s say, on a certain Thurs, did you assign the (Wed evening-Thur AM) sleep data or the (Thurs evening-Fri AM) sleep data for that day?
  • Curious to see what temporal patterns might exist with respect to cluser & inter-cluster relationships. i.e., do you tend to see more Cluster D days after Cluster C days OR do Cluster D days tend to run in batches (~ 3-5 days)?
  • I assume you used diaries/journaling to record your menstrual days?

Thanks so much!

Hi.
Thank you for your interest.
For the sequence: for a Thursday daytime data, the sleep would be Wed evening-Thur AM. This is the convention followed by the Oura output… so sleep preceding the day’s profile.
The temporal pattern would be very interesting! I haven’t looked at that in detail… but that is definitely something to check! Thanks for the suggestion.
And yeah, the menstrual days was from calendar records. I tried to get that data from my Garmin but there was no easy download option. So I simply relied on calendar annotations.

BTW, I think JMP generates the SAS code that details your analyses (to address your point about transparency)? Or am I mistaken?

Of course, SAS is not free, so doesn’t really address the paywall problem