Personal analytics

Quantified Self QS Open Forum
1

Greetings. Are you interested in learning more about how to use statistical science to confidently handle a data set? How about a Coursera class on a from-scratch re-envisioning of basic statistics, custom-built for self-trackers? If so, please see my guest post on MeasuredMe’s slick blog for more details, and get in touch with me (here, or adabney at stat.tamu.edu). There will be a preliminary Skype chat coming up sometime in the next couple of weeks to discuss, plan, scheme. Cheers. - Alan

2

3

Hi Alan,

I just sent you an e-mail indicating my interest.

Br,

Kouris

4

Thanks, Eric, heh heh. Possible date / time for initial Skype chat is Thursday June 6 mid- to late-morning CDT (Texas time), in case you’d like to join in; I’m adabney on Skype. The purpose of the Skype would be to brainstorm on stuff like: what level of mathematical detail to use in the class, what software to use in the class (for making graphs, computing stuff), what example data sets to use in class, what kinds of assessments to build into the class, when to launch, etc.

Thanks, also, Kouris. I’ll reply to your email momentarily.

For reference, and to help whet interest, I’ve prepared a representative example of what we could do in the course, attached and available here.

5

Any updates?

6

I’d be interested in this. Updates?

7

I’d be happy to contribute to the curriculum design.

The example given above implies a big problem in self-analytics is trying to separate what’s an important relationship or trend from what could just be randomness. This is a fundamental problem in stats of course. That said, there are lots of ways of going about it. Simulation, as demonstrated in this example, is a nice intuitive way. Of course, there should be discussion of experimental design–no use in going through the trouble of collecting the data if you can tell in advance there won’t be enough signal to answer the question you want answered.

Stats is a broad field, so the goal should not be mastering stats. What are the kinds of questions people want to ask about themselves? That’s where to start. From there we can see what analytical tools can be applied.