The correlation feature introduced in 2008 as relationships was poorly named when it launched. Today, we’re remedying that by renaming it “correlations” to better reflect what it actually is. We’ve also tweaked the algorithm to make it more accurate. It now uses the last 180 days worth of data, and requires as least 14 days to compute a meaningful correlation.
There are more enhancements in store, so please stay tuned!
We’ve moved servers and are ready for action again, hopefully faster than before! Sorry for the delay… it took longer than we would’ve liked.
Look for updates in the coming weeks.
It’s been awhile since I’ve posted (understatement), but I thought I’d share this article I came across on the Wall Street Journal: Happy? Statisticians Aren’t Buying It. Perhaps someday lifemetric can be a platform that statisticians can use for this type of thing?
With the help of the awesome New Relic RPM app, we were able to squeeze some more performance out of several areas of the site. I’m sure we’ll find some more hotspots and take care of them as soon as we can.
I’m excited to announce a brand new feature on lifemetric: relationships. Relationships are the calculated similarities from one metric to another. They are found on the right sidebar of your home page and user pages:
Let me take a moment to explain these results:
Generally the results can be interpreted as such:
| Coefficient | Interpretation |
|---|---|
| -1.0 to -0.7 | strong negative association |
| -0.7 to -0.3 | weak negative association |
| -0.3 to +0.3 | little or no association (filtered from view) |
| +0.3 to +0.7 | weak positive association |
| +0.7 to +1.0 | strong positive association |
We’re working on making these results a little bit easier to interpret at a glance, but the inital results have been quite intriguing.
(For the statistically-inclined, these are product-moment_correlation_coefficient">Pearson product-moment correlation coefficients, calculated from each pairs of metrics you are subscribed to and update on a regular basis.)_
How do you stack up against the world? Against your self? A couple new graph features might be able to show you a little better:
You can now see further into your past, up to one full year. In addition you can flip on your averages for the period you are viewing and the world’s averages.
To access these new graphs, click the graph on your home page.
The stress metric’s high and low endpoints have been switched. 0 now represents less stress while 10 represents high stress.
All previous entries have been switched accordingly.
Grouping our data by weekday provides some interesting results:
What do I take away from these results? The differences are subtle, but I certainly can tell that members tend to be happier on Fridays, which is not much of a shock given the history and normally relaxed attitude of the day. I also found it interesting that Wednesday was the low point of the week for the majority of members, not Monday as I would’ve thought.
Continue reading...lifemetric users that have forgotten their passwords can now request a new one be e-mailed to them, via the Reset Password page. If you have forgotten your password, enter your e-mail address and click Reset Password. A message will be sent to your e-mail address within moments with a new password that you can use to log back into lifemetric.
After you’ve logged in, you can change your password to something you prefer by clicking Profile.
Oh, and welcome back!
To prepare for our rapid growth, we’ve scaled up to two dedicated servers for lifemetric. It shouldn’t be a problem for anyone, just all the benefits of a speed increase.
Those with lagging DNS will still hit the old server, which will be turned off in a few days after the straggling DNS have caught up.