Ask HN: Please seed my survey to raise awareness about obesity | Hacker News Reader

Ask HN: Please seed my survey to raise awareness about obesity

obesitynearme.com

18 pointskbrower30 comments

I don't know if you're interested in statistics from outside the US, but if so, you should offer the ability to enter height in meters (or cm) and weight in kg.

Also, the A8 in "Paris, A8, FR" makes no sense, and it reminds me of those auto-geolocation ads that nonsensically say "Meet women in Paris, A8 now!" You should probably just drop the region altogether for most countries.

Finally, I don't think you should show any aggregate data until you get at least 4-5 responses, as you're leaking individual info. For example, just from what I've posted in this comment, anyone can deduct my height and weight from the "stats" on the site, which is not what I expected from filling out a survey (in this case I don't really care, but I'm sure others would).

It's true, Chrome is the lightest browser: http://obesitynearme.com/browser/

And Vista is indeed significantly heavier: http://obesitynearme.com/os/

Also people from reddit are apparently fat: http://obesitynearme.com/referrer/

(All results correct at time of publication)

On an individual basis waist-hip ratio http://en.wikipedia.org/wiki/Waist-hip_ratio is a better metric. BMI is a tool for populations as a whole.

Got a little bug with location... For some reason, it thinks "Montreal, Canada" is located in Ontario, so I kind of warped your results without knowing.

Also, and I know this is probably not intentional, but most people outside the United States will cringe when they see "and the rest of the world" with only a map of the states on the page.

Other than that, I liked the results page and all the statistics it provided!

These datapoints are a little scary:

- Your BMI is 26, which means you are Overweight

- Your BMI is lower than 42% of people who submitted this survey

- Your BMI is lower than 63% of 27 year olds

- Your BMI is lower than 39% of people from news.ycombinator.com

Sounds as if the people sampled skew young and overweight.

Tangentially, is there a better metric for something like this besides BMI? I recall some people having issue with BMI as a measurement, since it doesn't account for what your weight is composed of (fat vs muscle content).

edit: for anyone else interested in the shortcomings of BMI, wikipedia seems to have a good summary: http://en.wikipedia.org/wiki/Body_mass_index#Limitations_and...

its the best easy measurement. if you just want to track your own weight loss (and not compare to others), you can take waist measurements once you reach the point of gaining weight while still losing fat.

you can measure body fat % with calipers, but you need to go to a doctor with experience to get a reading with reasonable margin of error. http://en.wikipedia.org/wiki/Caliper#Use

http://en.wikipedia.org/wiki/Waist-hip_ratio is more useful for most of the obesity risk factors.

This is pretty cool. However, I recommend not ranking people in a weight class that equals their weight (IE You are % of people who weigh x lbs) because here will not be enough samples at a specific weight. Try using intervals of 5 lbs.

Good idea, I will make it work with 5lb intervals

Getting the captcha wrong clears all the fields. That's pretty cruel, usability-wise.

I will fix this

Another Kohana user I see :D

I caused an error :(

http://obesitynearme.com/city/Sleaford/H7/GB/

Yeah, Kohana is great! Thanks for catching that error. It has been fixed.

If this is to raise awareness, you might want to reword the stats. I found them slightly hard to parse.

"Your BMI is lower than 37% of people who submitted this survey."

This might be better written as, "You are fatter than 60% of people who submitted this survey." Ouch.

Also, if obesity is an epidemic then over time won't these results normalize? So I'm fatter than I should be but so is everyone else, so I don't look so bad in these results.

Can you get rid of the global letter-spacing? It’s OK for some titles, but it kills the browser’s built-in font kerning and makes everything harder to read.

Nifty. One suggestion: let me click on states in the image of the US instead of having to click on the image and then find the state in the list.

"Your BMI is lower than 0% of people from IL"

I'm a fat bastard!

I'm showing up as Windows 2003, but I'm on Windows XP 64-bit.

Yeah... I'm the fattest person in New York so far...

Your BMI is lower than 100% of people who submitted this survey

Is this your way of saying I'm a wimpy nerd? :)

My location was auto-populated as "Middletown, RI, US" - I manually changed it to be correct (Chicago, IL, US), but when I submitted the results page showed my % in terms of Middletown and in terms of RI. I was confused as to whether it was comparing me to RI for a reason, or if my Chicago input didn't stick.

Do you plan on only having data for the US? If not, have you considered adding the ability to enter data in metric?

I have just considered it, and I will add it.

Looks like there are a few other Canadians trying this out. I would definitely pass it on to friends if it showed a provincial break down as well.

I can't seem to select any location from England, UK. Typing "England" gives me various options in the US and europe. Typing "London" gives me options in south africa and US. "london, england" returns no results. Similar issues trying my small town or county.

My small UK town seemed to work - try a nearby city.

The app seems to use United Kingdom btw, you may have morel luck trying that.

I typed in my location as London, England but on the 'Your Results' page my location had been changed to Ilford K8, GB. It's very difficult to get it to show me the option for London, United Kingdom.

"Division by zero" error at http://obesitynearme.com/country/DE/

> Your BMI is lower than 97% of people who submitted this survey

And I'm much bigger than a few years ago...

And what use is the graph 'Obesity by Operating System'?

... at the risk of getting downvoted for low-brow digg-like links:

http://techcrunch.com/2008/09/09/are-linux-programmers-getti...

Reply on news.ycombinator.com