I began analyzing life-logging devices and apps over 2 years ago but neglected to write it up properly, so this post is a step toward addressing that. You can find the criteria I used to evaluate the devices and apps along with an explanation of how I developed the criteria on the Devices page. So far I have written up the analysis of visual life-logging, activity monitors and sleep monitors. I also submited a paper (2013-CHI-PI-Workshop-Price-Submitted) with a very concise write up of the study I did last year to the CHI 2013 Workshop on Personal Informatics.
I was disappointed at the really poor accuracy my participants found for sleep monitoring with the FitBit. My pilot participants had a rough correlation with their Zeo Sleep Monitor scores, but nearly all the participants in my study had very high sleep interruption scores which were very suspicious. Now that some of the smartphone sleep monitoring apps have more open access to their data I have been playing with them to see how well their scores correlate with Zeo or even my own subjective experience. If they even work for some people to some degree then they would be the least invasive, least effortful (virtually no charge/sync effort) logging method. Failing that, it remains to test more recent devices like the new FitBit Flex or the Lark. It may be that I was unlucky in my choice of particpants (there were only 7) and actigraphy is a poor measure for them.