Hacker News new | past | comments | ask | show | jobs | submit login
How the Napa Earthquake Affected Bay Area Sleepers (jawbone.com)
188 points by ismavis on Aug 25, 2014 | hide | past | web | favorite | 108 comments

The fact that Jawbone has this information on their users (yes, I know they say it's aggregated and "anonymised" data) is pretty scary to me. In my opinion, this data should only be visible to each user, and no one else. I really doubt Jawbone users are aware that their UP devices are sending data to third-parties... it feels like 21st century shackles.

Comments like these make me think that the pro-privacy crowd is officially a cult. I just don't see the argument about anonymized data that you can opt-out of being such a threat to the end user. I love these types of analysis, especially the ones about dating from OK Cupid. Tivo used to reveal all sorts of interesting things about viewership that was very enlightening. I'd hate to lose this data just because a minority think its a threat to their civil rights. There's a reason why no one outside of reddit and HN knows what the EFF is or cares about their mission. It doesn't fit in with most people's politics.

I really think the privacy advocates are holding us back at this point. They refuse to make reasonable concessions and while they think they're fighting the good fight, they've just marginalized themselves to the point where only a tiny extremist minority is engaged because moderate voices are shouted down as "selling out to the man" or the NSA or whoever the villain of the day is.

Its also sad to see a arguably progressive community like HN fall for this stuff. Not only are we doing this stuff all the time, the current fad of health devices, smart watches, etc are going to make this a lot more common. We're like cavemen throwing spears at an enemy with machine guns. Its a fight we've long lost. I think its better for us to focus on privacy issues that are legitimately victimizing instead of railing against anything and everything that kinda sorta rubs us the wrong way.

If anything, the cult of privacy is hurting us because its created a "boy who cried wolf" situation which is now a political liability. How am I supposed to tell Joe Midwest Voter that X is harmful when he keeps hearing about how tracking cookies or Tivo stats or iPhone policies are harmful and didn't give two shits about those either.

There is no such thing as completely anonymized data. And in this case there releasing information without the consent of their users. Sure, you might think that it's reasonable in this case but the important thing is users are not making that decision.

So, as long as companies feel they can do anything they feel like with the data your only option is to completely opt out. Which means companies have no reason not to release your data as everyone assumes they will anyway.

This isn't a question I have an answer to, but isn't it likely that users relinquish their right to data privacy when they sign a privacy policy and/or terms of use?

But they're probably not tremendously informed about this.

In clinical research it's considered unethical to take uninformed consent and a great deal of effort goes into educating participants on the risks and benefits of their contribution.

Such stuff is fantasy for modern EULAs though, which makes it sketchy when these entities do (public) research.

You're extrapolating from one comment the viewpoints of everyone who cares about privacy? That's quite ridiculous. As another commenter has mentioned, de-anoymising data isn't that difficult [1].

Why should privacy advocates 'make reasonable concessions'? What are they holding us back from? What concessions is the other side making? Who exactly is on each of these sides anyway? Your comment isn't particularly clear on these points nor on why the opposite view is any better (other than we get nice graphs occasionally). There are cogent arguments to be made on either side here but you've made none.

> "I'd hate to lose this data just because a minority think its a threat to their civil rights."

Minorities matter and it's ironic that you bring up civil rights. There are plenty of example from history so please do look them up.

[1] I'm not saying it's possible for the graphs in the article, but if one had access to the (anonymised) data that produced them, it may be possible to combine with other data and ultimately identify individual users.

I'm a Jawbone UP user and I have absolutely no problem with this anonymised data being used like this. It can make not only the product better but could also lead to useful research around sleep and health.

If I want out then Jawbone has a handy 'remove all my data' option.

People share all sorts of data in all sorts of places (as others have pointed out in this thread), and it is their own responsibility to read the T&Cs of those services.

I agree that data privacy is important for those who want theirs protected. But in this instance, people have chosen to wear their UP and have T&Cs which notify them of the usage. If they later change their mind then they have the option to have their data removed.

I can't see how any of that equates to 'shackles', but agree from some POVs it could be considered 'scary'.

> "... it is their own responsibility to read the T&Cs of those services."

While some people may believe/feel this is reasonable, it has rapidly become impractical. Have you read the T&CS for all the services you use? Any idea how many pages of text that comes to? Did you also know that the reading level of most T&Cs requires college-level education (just to comprehend)?

My point is that it is already impractical (& will become more so) to suggest that T&Cs can make people aware of anything.

I used to read them, then sift through them. Then I gave up. My defence would go along the lines of: you don't seriously think that any end user would read that, would you? (Exceptions for APIs etc.)

Same goes for legal disclaimers in email footers, cookie warnings, warning stickers etc. Why are we allowing this to happen?

(My solution: "Yes, you can sue them because the coffee was hot (etc) and there was no sticker on it, you just have to be prepared to give up your voting rights and your drivers license in the progress. Continue Yes/No?")

As an aside, the "coffee was hot" lawsuit wasn't as frivolous as people make it sound. See Liebeck v. McDonald's Restaurants for details.

Anytime someone says that lawsuit was frivolous, I lose a bit of respect for them.

Their privacy statement is short and clear:


~1400 words.

I liked this section:

When you use the Jawbone Companion app we connect with and upload to our servers the address book and calendar on your device along with other data that we normally collect through our mobile apps described in the app usage data section below.

I wonder how calendars are used.

I can't tell if you were being serious. Assuming you were, then 1400 words, presented as a wall of text, is an anathema to the short attention spans of many users. In addition, that 1400 words is a subset of around 5400 words for the Terms and use of software/services.

We can debate whether people should be educating themselves about such terms but my point is (still) that it's impractical in the long term as we want to use more services/devices. The only way out of this that makes sense to me is that people begin owning their data and (digital 'exhaust'). I'm working on FOSS tools to enable this (see my profile).

Yes, I was serious. It's a $80 dollar device that you wear a significant amount of time. In that context, 15 or 20 minutes spent understanding what it does is not onerous.

Note the comment I made earlier in another part of the thread where I said I pretty much refuse to use device+service combos like the one here.

In that case, I'm curious whether you use a smart phone (iOS/Android) and whether you've read all those documents too. That's typically a much higher-value product.

Also, recall that I said that a certain level of knowledge/understanding is required to even understand the basics.

> "I pretty much refuse to use device+service combos like the one here."

Could you elaborate why? Onerous T&Cs? No control of data? Don't care about such analytics?

My problem with how things are developing is that we end up in a world where either you resign yourself to the fact that you're constantly being data-mined or you become a digital hermit. There's very little that allows you to engage and retain control (without becoming a sysadmin).

I have an Android tablet, but not a smart phone. I'm not hugely careful with what I stick on there, but I think I've only shared my contacts with stuff like Hangouts and Skype (where the explicit purpose of the app is to contact people). I increasingly leave my cell phone sitting at home (mostly because I don't care about being reachable, not because I'm worried about network location pings). I take it with me if I'm going to be away for more than a couple hours.

As far as not using analytics, it's a combination of not caring (as an example, I run (maybe I should say jog?) regularly but don't time myself) and distaste over the way the data is managed.

The fact that people's views of their privacy rights are subjective means that a lot more than 1400 words would be necessary to cover every single action that at least one human perceives to be a privacy violation.

Maybe there should be regulated permissions, sort of like the way an OS might have a dialog to request permissions for push notifications or location data.

Example, you can still include any terms you want but some items need to be individually agreed to in plain language, like "This product will report your location back to us"

Sort of like FDIC notices or nutrition facts.

That is a good point - it is quite difficult to read and interpret all these sets of T&Cs these days. Something should be done to make these more accessible / standardised.

However, that is a problem that is equivalent across all facets of all services people use, and I think is independent from this problem.

The likelihood is that most people don't read the T&Cs and most people don't care about their data being used like this. But if someone is concerned about a particular facet of a service (privacy/data usage) then there is a place they can look and relatively easily find out what they need before signing up.

What can be done about the wider problem of T&Cs and how they are unwieldily is a tough one.

> Have you read the T&CS for all the services you use?

No, but if I personally thought that practices which are extremely commonplace (like the collection of anonymous data) were privacy violations, I probably would read the T&Cs. Unfortunately, I would also be unable to use the vast majority of electronic products, including virtually all websites.

And has anyone seen the notice "This site uses cookies" and said "aw shucks"?

And if they do, do they also continue using the sites that don't have to tell you about their cookie policy?

I still believe it's reasonable because once you've read one, you've read them all. They all pretty much say the same thing. Jawbone doing this or Facebook experimenting with the feed should not come as a surprise to anyone who has skimmed over a TOS.

> "They all pretty much say the same thing."

This is not a safe approach when you're dealing with legal contracts, which is what T&Cs purport to be. Minor difference in wording can have very different intents/outcomes.

Having said this, they pretty much give carte blanche to the service provider to do whatever they want -- so in that sense, they're all the same.. That doesn't make it right and it doesn't mean it can't/won't be contested at some point.

At first I thought something along these lines, then I remembered:

1. Google has "trends" when things happen, and "Earthquake" was most certainly googled near/at the time it happened. 2. The phone company has records, which many people texted/called near/at the time it happened. 3. 911 has records and calls were made near/at the time it happened.

and so on…

So there are private entities and government that have data around the event that can correlate behavior and situations, so it shouldn't seem scary to me… it's just should I have better access to the data? And if I do, what can I tell from the correlation?

I think it also has to do with what content is being recorded.

Intimate sleep data is a bit different than knowing I called 911 for an emergency situation.

Are you saying that the sleep data is more personal than calling 911? I have to question that, because you could call 911 for all sorts of private situations that you wouldn't want publicly known. So no, it's not really different.

Especially in this case. Saying something like, "I couldn't go back to sleep after that earthquake woke me up" is the kind of thing many people would share on twitter, facebook, or around the water cooler with coworkers.

It's subjective. I would consider a call I made to 911 to be a lot more personally sensitive than my sleep patterns.

Quote from [1] Jawbone Privacy page: "We may share your Information with third parties to provide services on our behalf such as to process payments, or to store information collected through our site, app, and services." Does this statement allows them to use users data for this purpose?

[1] https://jawbone.com/legal/privacy

The Use of Information section includes press releases although it is not counted in the sharing section. I also see no commitment to anonymising the data.

Jawbone uses your information to provide and deliver products and services you request; personalize your experience and customize content, marketing, and recommendations; respond to requests that you make, provide you with customer service, send you press releases, general informative announcements, and promotional information about us or our products, or to aid us in serving you better; and protect, investigate, and deter against fraudulent, unauthorized, or illegal activity.

Well, we the general web audience aren't providing any services to UP wearers on behalf of Jawbone, and they're not giving us the data to store (that is, they're not asking us to store it), so that quote doesn't allow such a sharing of an UP wearer's data. However, they might be willing to argue that sharing aggregated data isn't the same thing as sharing your personal data; it wouldn't be a totally ridiculous idea.

Oh, come on. I'm getting a little sick of how far the pendulum has swung in this privacy conversation.

I get the desire for privacy, but if every time people have data about them used we get this complaint, no one's going to listen when there's a real violation of privacy.

You don't need a sleep monitor to function in the 21st century.

If you try to take a Jawbone off, I believe it will take a few seconds.

I dislike these data collection and analysis as a service gadgets enough that I'm not using any of them, but I think you could be a little more considered in your criticism of them.

As an UP user, I think Jawbone make it pretty obvious when using their app that you're sending data up to a server, and as others have pointed out, the EULA explains that this data can be used elsewhere.

The UP bands data processing happens server sided and is indicated by the UI. Was there an indication here that a third party was involved (the post indicates its their internal data team)? Otherwise I think everyone has generally accepted internal company use of data (critical part of growth) let alone when its aggregate and anonymous.

I look forward to NSA serving an FAA 702 order on Jawbone to track a non-U.S. user.

It would require that Jawbone perform a live intercept so FedGov agents can plant physical bugs or insert TURBINE-derived malware or attach a keystroke logger. The order would use existing legal authority to compel Jawbone to provide real-time data on the user’s precise geographical location and whether he or she is awake. Obviously if the target is at home and awake, the Feds would abort the black bag job.

The Google Play store listing shows that the Up app wants access to precise (GPS) and approximate (network) location, so the lat/long available for NSA perusal with compulsory process is likely pretty detailed. And the marketing literature on the Jawbone site uses “always-connected” and “real-time” language, which is perfect if you're in the real-time surveillance business.

The non-U.S. user targeted by FedGov may think he’s escaping surveillance because he’s using only local phone providers and local email providers. He’d be wrong; Jawbone is based in SF and subject to U.S. law… Another interesting use would be tracking physical movements of non-U.S. users and correlating them to see who’s meeting with whom; does Jawbone force SSL for its API usage? (The API web page only says “communication should be done over a secure channel,” not “must.”)

I’m joking above, of course, and I have no reason to believe the NSA has done this for a non-U.S. citizen or the FBI has for domestic surveillance. My point is to suggest that folks here at startups think in advance about how to respond to legal process, how to encrypt data in transit, and how to store user data in a way that minimizes your legal exposure and protects your users’ privacy.

Better to think about this early vs. after you get an early morning knock on the door with a very serious plainclothes guy with a badge serving you with some very serious papers demanding info. That happened to me in a different context years ago, but that’s a story for another time.

To me, there's a world of difference between sharing my personal commute between home & work, and using my drive to show traffic patterns.

Does anyone know if there's a device like this that DOES NOT require uploading all collected data to "the cloud" in order to analyze what I've collected?

It's pretty obvious that this could be done on client side without needing an internet connection. Or do all of these companies sacrifice usability and privacy in order to make more money?

First off, I have a Jawbone UP24 and love it. You know how they got that sleep data? I pushed the "start sleep" and "stop sleep" buttons, then they gave me an awesome visualization of how I slept last night. You know what you'd do if you didn't want them to have that data? Don't push those buttons.

How are they not aware the data is sent to 3rd parties? Isn't that a key feature of the device?

Total privacy invasion.

Am I the only one that's finding a bit upsetting the fact that a single overlooking identity could dispose of people's data like that? I know, I know: all the Up wearers have "enrolled" voluntarily, but I wonder if they really like to be part of a giant anonymous data science experiment. If you can do this with the users' data, you can certainly do the same for more profit-related purposes. Or maybe it's just me, sometimes I just can't ignore the scary within the scientifically cool.

Anybody who is willing to (A) Track their personal biological information, and (B) willing to upload this information to a third party, is utterly unconcerned about an aggregate graph like this.

Your meta point, though, that people should be concerned about this type of tracking is well founded - but, honestly, I'm probably a lot more concerned about how easy it is for my cellular provider to track where I am at any given instance with my Smart Phone - (or any cell phone for that matter) - as I don't recall ever "enrolling" to provide them with my tracking information.

Did they really dispose any of the users data though? It appears they displayed a chart.

That graphic, it could be argued, is based on an entirely new data set.

So while individual data sets may been used at some point behind the scenes, it's not like those data sets are the ones being disposed.

the issue is that they're collecting this data in the first place.

I cant help but wonder how much it would cost to get the sleep records of a competitor of mine - or if it's not for sale, how much the data would be worth on the black market if their servers got hacked.

I have doubts this information is being kept as securely as it should be, that's where my concern lies.

The whole point of the device & service they are selling is, they collect this data. Then you can log in and see it in the future.

Distributing is a different concern. But people pay jawbone to actually collect and store it, so they can view it later.

Well, if one is willing to one-up a fellow human like that, one is a bad person. But you don't need Jawbone data to do that, there are plenty other ways. Honestly, the problem in your example isn't the data.

one is a bad person

So what? Time and time again things that would make you a bad person are silently accepted as normal or even expected of a business.

You're correctly describing the state of the world, but my indirect point was that we should stop accepting such behaviour in business and shun people who do. It's stupid and wrong.

Interesting data. Some things that jumped out at me:

1. Apparently everyone in Santa Cruz goes to bed early and gets up early.

2. I would have expected bigger jumps on the hour. Maybe I'm the only one that sets my alarm on the hour though.

3. Why would there be dips in the AM? People hitting snooze perhaps?

4. How long do you have to moving for UP to think you're awake? How long do you have to be still for the UP to think you're asleep?

> 1. Apparently everyone in Santa Cruz goes to bed early and gets up early.

Probably some selection bias at work here. It's possible the type of person who's buying and wearing personal fitness monitors is also the type of person who gets up early.

> Maybe I'm the only one that sets my alarm on the hour

Probably my OCD coming out but I only set my alarm with minutes beginning 0-4 and ending with 8. I'm sure everyone has their own habits. On the hour seems so numerically boring!

If it was selection bias like that, the same would apply to people in other cities, so it wouldn't explain the difference in the graph. My guess is people in Santa Cruz on average have a longer commute.

It's going to be a selection bias of:

1. People who are fit (or want to be)

2. People who are technical (or are interested in quantified life in general)

Technical people who commute from SC to Silicon Valley have far less bus options and routes than SF (SC has three each way for Google which ends at 9:25). That forces a lot of the number 2 people awake earlier than they otherwise would be.

I think SC is probably the same, if not average getting up later than other Bay Area places, due to the large student population and short walk/bike commute for those who work in the town.

> Maybe I'm the only one that sets my alarm on the hour though.

I assume most of the UP-wearers will use the "smart alarm" function on their wristband that will wake them at the "optimal moment" in their sleep cycle.

> 1. Apparently everyone in Santa Cruz goes to bed early and gets up early.

What I'm guessing is that people in areas of higher population density stay up later and get up later on average. The 25-50 and 50-75 mile bands seem to contain the big urban centers, while the 0-25 and 75-100 bands are more the suburbs, small cities and rural areas.

I would expect that most people set their alarms in 5-minute increments, and dial in what minute they choose to meet the needs of their commute and morning routine. This forces variations away from a "sharp" time spike. Also, alarm clocks are almost never right on time; I would expect that most alarm clocks are more than a minute off. This factor would smooth out the 5 minute steps I'd expect to see in the data. The alarm clock inaccuracy may decrease over time as people move to using their smartphone to replace yet another device.

Since they don't say whether each line represents the same number of people, I would guess that an outlier line is just a smaller sample.

Did some work in the area a few years ago. Found out most of the people at the place I worked for got up at about 2am (!) because they wanted to be finished by lunch. At least in the summer. (disclaimer: this wasn't an air conditioned office though : )

It's unfortunate but almost all Quantified Self devices work like this. You have to go through the device provider website to see your data.

I wish there existed more engineer-friendly initiatives where you can install the data collection server, data store and analysis server on machines you control.

If these consumer-level devices are $100-200 a pop don't you think that would be prohibitively expensive for a decidedly smaller audience?

In response to antr and others talking about how this is "private" data, a few things:

1) Jawbone is a data company. You get what you want from them by giving them your data. Giving them your data is your choice, so if you are uncomfortable with it then you should not have a Jawbone. 2) What's interesting about this article is that it's in aggregate, across a statistically significant (so they say) number of people in these areas. They're not looking at individual stats, and I would bet would have a hard time tracing it back to a single person/name/etc, if they're even allowed to store it that way (not sure, but it would surprise me if they did).

If you don't like your data being stored somewhere, why are you even on HackerNews? You know your data is being stored here and being learned from.

> If you don't like your data being stored somewhere, why are you even on HackerNews?

having my sleeping records and a table of what times during the day that I shit is a lot more sensitive information than the trash I post on hacker news

Once again, if they have your sleeping records it's because you have chosen to give them that data. It's your problem then, not theirs.

Also, if you think others care about what time of day you take a shit, you have way too high of a view of yourself.

That is your personal opinion, and it's perfectly valid. I would recommend people with that opinion to not use a Jawbone UP.

Here I am, a resident of Melbourne, Australia, where we never get interesting earthquakes, visiting SF for the first time (staying in Belmont) and they don’t even wake me for the earthquake they evidently put on for my benefit.

Hey fellow belmont buddy! Just curious, about how old is the building that you were in when the earthquake hit?

It's just that Belmont's on bedrock. Take Ralston east toward the bay, out on the landfill a few blocks past Oracle, and you won't be sleeping through it. It shakes like Jello out here.

How useful do you find the sleep tracking to be? I've been using it for a couple of days now and I'm on the fence whether a simple movement tracker can really tell whether I'm in a deep or light sleep phase...

Think of it comparable the difference between "pop science" and "academic science" in book publishing. Its cool, and usually pretty informative, but largely just for fun or minor utility.

Who else finds this as creepy as hell.

Who else finds this cool as hell?

Me too! I think it is great we have progressed to the point where we can have this sort of insight. :)

Imagine when NSA starts to publish such charts. They have really COOL data.

Sure they do. And they could do a lot of good with this data.

Post-Snowden PSA: Remember, they already have it if they want it.

Creepy is the new luddism, I think. Seriously, what the hell is so "creepy" about someone knowing your sleep pattern?

Nobody is saying that the technology is creepy. It's the fact that it's not at all private that's creepy.

As Uncle Ben said, "With great power comes responsibility." It's an awesome thing to have the power of "big data" to analyze your life and make better decisions. But "big data" needs to be responsible in how they handle that data (and their users' trust).

Which individuals data was compromised in making this chart?

We don't know

That's the whole point. The fact that they have this data means that they could share it with whoever they want, whenever they want. And we would never know. So in this particular case, it's not a data issue - it's a trust issue.

The Jawbone device is built upon the idea of the company storing the information. It is implicitly implied that if you use the device your data is being stored somewhere because that is what you are paying for them to do

Sure - but that's not the issue that people have with this. The issue is that Jawbone decided to release personal data (aggregated, but still personal) to people who did not own the data.

Like I said above. . . it's not a data issue. It's a trust issue.

Since Jawbone apparently don't know how to post a simple image for old-school browsers: http://blogs-images.forbes.com/dandiamond/files/2014/08/Jawb...

And for those who've noted that the NSA are going to have all the cool data, after seeing this chart, I'm pretty sure they'll start tapping into Jawbone as well.

But this is not simple image. This is interactive chart. You can hover lines and get some info via tool tip. Apparently you are using quite old browser, if you don't mind asking you? Since D3js, the lib used to generate SVG, supports majority of browsers[1].

[1] https://github.com/mbostock/d3/wiki#browser--platform-suppor...

You are correct. I am using an old browser.

It does date from the present decade, however. It's not that old. I'm somewhat amused by the number of sites which fail to work properly (some at all) with it.

Imgur is entirely dead. Reddit I can read, but moderation and reply buttons don't work. I hit the beta rollout of a news site last night (apparently part of their A/B test) and found that among the features which failed to work was the "opt out of beta" button.

So... Update, maybe? I'm sure the web developers of those sites aren't "somewhat amused" by having to try to support your ancient browser.

D3 should work in any browser that supports SVG, which is every desktop browser but IE8 [1]. If you're still using IE8, some sites not working is going to be the least of your problems soon - Microsoft isn't patching it anymore.

[1]: http://caniuse.com/#feat=svg

The OS I'm using isn't eligible for an upgrade. So that would require hardware. That's not in the budget at present.

The browser isn't a Microsoft product.

Why are you complaining? Web developers are not obligated to support people who chose to browse the web through a broken interface.

Nor am I obligated to use systems and services which don't support minimal HTML interfaces. Graceful degradation is generally useful, particularly as most web spidering is still basic HTML (some now support aspects of JS as I understand). You're invisible to the world if you can't get indexed.

I've also found that the very rudimentary browsers included in cheap feature phones tend not to work particularly well with any measure of "modern" websites (though others do support these fairly well). While I don't suppose these are a large financial market, they're likely a large user market, as those phones are both cheap and exceptionally thrifty with battery life (weeks, not hours).

Given that this is graphical content, that's less a concern. But as I noted above: simply providing a fallback JPEG alt would have worked.

And noting an issue isn't the same as complaining. Actually, if you'll note, I fixed the problem, after a fashion, by providing a link to alternate and accessible content.

But thanks for your depth of understanding, empathy, and sympathy.

A map version of this (with contour lines for percentage awoken) would be nice. This would also eliminate the small differences in the time people in different areas seem to go to sleep and rise.

Very cool!

The text says that 93% of the Napa area woke up, but the graphs shows 74%. I'm assuming the graph is off?

The graph is showing how many people are awake. The 93% number reflects the fact that some people were already awake before the quake.

Then according to the graph, 59/85 were awoken from sleep, which is about 71%.

59 awoken = 74 (awake at quake) - 15 (awake before the quake)

85 = 100 - 15 (awake before quake)

Maybe they are adjusting for other things that are not shown on the graph.

Good point. I wonder if they are smoothing the data, which would make the peaks less peaky (which brings up another question, what is the time resolution for UP's ability to measure wakefulness?).

Never mind. The graph is showing areas based on 25 mile radius, and the 93% is for the 15 mile radius.

Maybe more interesting is that Santa Cruz/Modesto people are early to bed/early to rise.

A number of people live in (beautiful!) Santa Cruz but commute over the hill to work in San Jose or elsewhere in Silicon Valley. It's a long commute on a treacherous but crowded mountain highway, so it's no surprise that they need to wake up early to beat traffic.

Link currently leads to a 404.

Serious question:

I wonder if they can track how many people die per year and differentiate between type (car, earthquake etc.).

That's not the point. Jawbone (and similar devices such as FitBits and Nike FuelBands) can detect whether the user is sleeping or not; this post asserts that the earthquake is the primary cause of the spike in sleep disruption at 3AM, which is a valid and supported hypothesis.

As far as I know, these devices cannot detect the cause of death.

But the device detects movement doesn't it? Is it not able to detect seismic activity or any other force? So if a person was in a car accident or in an earthquake that would be something that can be measured.

There are now devices that can tell what exercise you are doing just by being on your wrist.

I don't know much about this device but I guess it doesn't do continuous synchronization. I believe the user need to initiate it which won't happen if he die.

"All results are statistically significant."

Isn't this an overly optimistic claim? Surely people using a sleep tracking device do not form an unbiased sample of the general population when it comes to sleep patterns?

Nobody said it was a study on the general population, only on the UP wearers population.

OTOH, if their sample size is the same as the population size, their data is by definition statistically significant.

Not even that. If the sample size in general is extremely large (irrespective of its proportion to the population size), than the results are statistically significant.

A sample size in the thousands counts as "extremely large."

Surely the sample, however large, must follow the same distribution than the population distribution? If you sample 1000 insomniacs, how do you get anything significant?

My remark was just that, the sample bias has an impact on the significance of the results. I do not know if in actuality people using sleep monitors form a different distribution than the general population regarding sleep patterns, but it seems likely.

The title of the article mentions "Bay Area sleepers", so it would seem to me that they do form conclusions about the general population, not just Jawbone wearers.

I think that most people greatly overestimate the sample size needed to achieve statistical significance.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact