about:profile – Analyzing Data in Firefox

[Edward Lee]

The Prospector team has a proof-of-concept experiment about:profile,
which analyzes data stored locally to an instance of Firefox, to start
a conversation about different ways to analyze the data with the goal
of improving Firefox while supporting Mozilla's principles of openness
and user privacy.

https://blog.mozilla.org/labs/2012/10/about-profile-analyzing-data-in-firefox/

Please check it out and provide some suggestions!

Ed

[B.J. Herbison]

The top sections seem reasonable, but some of the demographics are way off.

(I suspect that green is for likely and red for unlikely because that interpretation seems a better explanation, but either way it doesn't reflect reality.)

[Edward Lee]

On Sat, Oct 6, 2012 at 5:07 PM, B.J. Herbison <bjher...@gmail.com> wrote:
> top sections seem reasonable, but some of the demographics are way off.

Are any of the demographics correct? You can click on each of the bars
to see which sites contributed positive/negative to the value. We only
packaged 500 sites for demographics and 5000 for categories, so it's
expected that the results won't be accurate for everyone.

Curious, do you visit many "non-US" sites? The packaged data doesn't
cover many regions. Do you have suggestions for how to improve the
demographics calculation?

Ed

[B.J. Herbison]

> Are any of the demographics correct?

Now that you mention it, only one. The Male/Female selectors are even so I count that a's "no guess" and not wrong, one is correct, and the other three are solidly wrong.

I was going to say that the guess of "Work" is reasonable as I do some fairly technical research for my own projects, but then I clicked and saw linkedin.com as a reason for assuming I was at work. I never visit LinkedIn from an employers computer, it's my personal network and account. (And two of the web sites listed as reasons against me being at work were sites I regularly visit from my employer's computer.)

> Curious, do you visit many "non-US" sites?

No, mostly US based. I did see some non-US sites contributing to the demographic scoring: some supported my actual demographics, and other pointed you in the wrong direction.

Have you considered setting up an anonymous survey where people can report what about:profile says and compare it with their real demographics?