10 ms·

Fitbit, why can’t I have my data?

211 points by KC8ZKF 14y ago

darklajid 14y ago
My pet peeve: I wanted to know what fitbit is, but the site redirects me to fitbit.com/de and I closed that site. How can they be open about data, if they have obviously no clue about open standards, like http accept headers? Worse, there's no way to visit the (official?) fitbit.com site even after that idiotic redirect. I can't access an english version, period. Stupid.
- ig1 14y ago
  There's a logical flaw in your reasoning "doesn't http accept headers" doesn't imply anything about their understanding of open standards. While it's perfectly understandable that you might be upset that their site for Germany doesn't offer english language as an option, that's clearly a completely separate issue from how they handle user data.
  - darklajid 14y ago
    You're correct. I let my anger about ignoring my explicit preferences, which I expressed in a standard way, lead me to believe that the company doesn't care. Obviously that's annoyance speaking and speculating based on me being tired with the constant 'Hey, I know better what you want' attitude (Looking at you, Google/Blogger) - I don't even know a single thing about their product (which is why I went to the site in the first place). Sorry about that. I stand by my point about this sort of redirection being a telltale sign of a flawed web site and lack of respect for user preferences. What else they can or cannot do, I have no clue about and cannot judge.
    - ig1 14y ago
      Given that you're in Germany and they don't have an english version of their Germany website available their two options would be either to send you to the German language website for Germany (which they do) or alternatively send you to the US or UK website which are in english. I'm not sure it's obvious that sending someone in Germany to the UK/US website (where presumably they'll be unable to buy the product) just because their browser is set to English is the better solution. While for sites like Blogger language is obviously more important than country, for companies which are country-localized (i.e they treat different countries differently for shipping, taxes, legal, operations, etc.) I would guess that it makes more sense to send you to local country version. Imagine you were using say a dating site or a takeaway site, you would find it equally frustrating if you were routed away from a local language site to a US specific site just because that's what your language preferences were set to.
      - darklajid 14y ago
        I'm not sure why you're conflating a) content and localization b) the region I'd like the product to ship to, if at all Why is the site different for different countries, ignoring 'translation'? I haven't thought it through, maybe, but I cannot come up with any decent reason for a 'German' site that isn't just the 'US' site in a different language. In that case, please (dear website) listen to what I'm asking for. If I go to fitbit.com I expect to get the very same thing someone in the US receives. I'd like to talk about the very same site. I don't want someone to redirect me to a localized thing. And certainly not without giving me the opportunity to say 'Yeah, no. That was stupid. I really wanted the original version, silly'. Same thing: If I go to www.google.com, I want to end up at www.google.com, not www.google.de. If I visit a random post on Blogger, chances are everything content is in English. Except for the 'helpful' Blogger toolbar and whatnot, that are coming up in German, because hey that's where we figured out you're coming from. Lived in Israel for a year, got a Hebrew toolbar, google.il (and I'd like to know what fitbits would've done there). German vs. English is one thing: I can read both, I just explicitly (url, domain, accept headers) ask for the latter. English vs. Hebrew is another: I cannot read the latter, even if I happen to be - yay for geolocation - in the one state that represents the Hebrew language. Imagine you were using say a dating site or a takeaway site, you would find it equally frustrating if you were routed away from a local language site to a US specific site Right. Don't send me anywhere if I navigate to example.com, even if I ask for de_DE. Offer a translated version, if you can. Otherwise drop a small (German?) link on your .com, saying 'We noticed you explicitly ask for German content. We got a country specific site right here -> example.de' A dating site would allow me to register and state my country of origin or interest (which might be Germany, even if I live in Tel Aviv at that time). A takeaway site is really a weird example. www.pizza.de is available in German only for all I can tell and won't redirect me to a random US site because I ask the server to please return en_US or en localized content, _if possible_. So, for me this whole 'automagic-we-know-it-best' translation/redirection thing is broken by design. It was a constant hassle in the past and just seems to catch on. Which is why I'm pointing it out when I can. I'm sorry for the thread-jacking. Thanks for the exchange so far.
gregcohn 14y ago
Good post. Are are any of the similar devices more open, in terms of both data accessibility and format?
- jrockway 14y ago
  If a FitBit is just a pedometer, I have an Omron HJ-720IT that syncs with my Linux box (uploading a week's worth of hourly step counts). It's $30 on Amazon.
  - thisone 14y ago
    The cheaper one is a pedometer, the expensive one tries to do things like track your sleeping patterns and hooks into their own make of scale.
mavlee 14y ago
I actually ran into the same problem recently, when trying to get my minute by minute data. I even emailed their support team, but no luck. I wonder if there's an easy way to scrape their flash charts on the dashboard...
- MichaelApproved 14y ago
  Without using the website, I suspect you can use something like Fiddler to see what HTTP requests are being made. Chances are the charts are making an XML or JSON request to their servers. If you're lucky, the URL requests will be in an easy to reverse engineer format so you could easily dump all the data you need by adjusting the URL.
  - bloaf 14y ago
    Why not try to intercept the data as it comes through the USB port instead?
    - Thrymr 14y ago
      Or try liftbit as noted above: https://github.com/openyou/libfitbit https://github.com/openyou/libfitbit
      - simcop2387 14y ago
        libfitbit unfortunately doesn't currently work with the fitbit one, for the older devices it's apparently fine but the one changes how it gets talked to entirely. I've attempted a bit of work with getting something talking to it but it's not an easy thing to do. Their bluetooth dongle presents itself as an HID device and I think hides a lot of the details away from you if you wanted to talk to it over actual bluetooth. I'd love to get it to sync with both their site and to be able to capture the information myself to store.
- fla 14y ago
  The flash content must get the data from somewhere, probably in XML, certainly over HTTP. Open Wireshark and capture a session to known where it came from. Then write a program that replicates this behavior and save the data.
harold 14y ago
This project hasn't been updated in a while but might be a good starting point for a roll-your-own solution: https://github.com/loghound/Fitbit-for-Google-App-Script https://github.com/loghound/Fitbit-for-Google-App-Script
nodata 14y ago
Just wait until a large health insurance company buys Fitbit (or Withings, or Garmin) - they will have your data.
- jrockway 14y ago
  Just wait until a large health insurance company breaks into your house at night and installs cameras - they will have your data.
- alexkus 14y ago
  I have various Garmins (Forerunner 405 for running and Edge 705 for cycling). None of my data goes to Garmin as I just use Garmin Training Centre and don't upload anything to Garmin Connect.
  - npsimons 14y ago
    Much as I dislike Garmin's lack of support for Linux, I have to second this. My Forerunner 405, eTrex Vista and Vista HCx all download to my computer, with the option (not forced) of uploading. I have full access to .GPX files that I can use with any number of pieces of software (including Google Earth and Maps). Also, I've been pretty happy with the Zephyr HxM hooked up to my N900, and the wife has a Zephyr as well that works well with a Samsung Galaxy III and Google tracks.
    - quantumstate 14y ago
      It case you haven't come across it I have found that garmintools and pytrainer work very well with my forerunner 305. I have never used anything else.
MichaelApproved 14y ago
"I guess it is true, if you aren’t paying for it, you are the product." But according to the author, even the $50 option doesn't let you download the data so it has nothing to do with you being the product.
- vertis 14y ago
  More than that, you have to pay for the device in the first place, so on that front you're absolutely the customer.
  - jrockway 14y ago
    Indeed. Everyone likes to think missing features are some scam that the company that owns the project is perpetrating upon their users, but in reality they probably want to lock you in for the sake of locking you in. They don't know why, but making the data open would be extra work, would help their competitors, and might hurt their chances of being bought. They can always be more open later, but they can't ever un-open. So they're closed. Don't buy their products. That's what I did.
    - angdis 14y ago
      I agree, there are other reasons for not providing the raw data. For the vast majority of people raw-data just isn't something that is even remotely usable. Even if one is an expert with data analysis, IMHO, it will require a lot of work to process and interpret it correctly. It may even require a deep knowledge about the device and all its quirks-- we're talking about data derived from a cheap accelerometer here. In the end, it is hard to see how opening up the raw data so a few hackers can take a crack at it is advantageous to fitbit or even consumers in general.
randomchars 14y ago
If you live in the EU you can send them a data access request, which means they have to provide you with all your personally identifiable information that they keep, in this case that means pretty much everything. I hate it when companies think that they can get away with making you pay for your data.
- dawson 14y ago
  I have always been curious about this, from a healthcare perspective. For example, we encrypt PID information one way (the patient has their own unique data access key, which is then encrypted with their password + salt and some 'other stuff'), so we couldn't hand over any information if a data access request was made, even if we wanted to as we don't have access to that information ourselves!
  - hayksaakian 14y ago
    You could hand over the encrypted data, the usefulness of such data being low however.
    - dawson 14y ago
      The issue is that the PID is all encrypted, I wouldn't know if we have a John Doe on our system, we can't see anything to verify that user exists or that they're the appropriate owner of said data. We do have an email (which is also encrypted at a system level, which in theory we could access), but verifying an access request based on an email address? I don't know. The way we solve the problem is by making all the data available to the user once they have authenticated, as long as you can login to our website, all your data is available to you, but if you made a direct data access request to us, like a FOI to our office by letter (the typical way it's done), not much I think we could do.
      - thisone 14y ago
        Your IG team will know. From my understanding, from the yearly IG briefings I used to have, you would be able to provide that requester with information on how to obtain their personal access, and point out to them where they can find the information they have requested. Information needs to be made accessible, you don't necessarily need to actually provide printouts. For example, FOI requests from news agencies often cover the same topics. This information can and often is posted publicly on the organisation's website and the FOI responses refer the requester to those links.
yarrel 14y ago
There's libfitbit - https://github.com/openyou/libfitbit https://github.com/openyou/libfitbit
- rb2k_ 14y ago
  The new fitbit can sync using the iphone's bluetooth 4 connection. This makes syncing a passive rather than an active thing which was the reason I bought the fitbit one even though I had the original one.
biturd 14y ago
Why do we need an actual hardware device to get to this data? Don't iPhones and Androids and others have all the internals needed to get the same data, and more, such as altitude, so you can check the incline you are against etc.
- heyitsnick 14y ago
  I think for lots of people, they don't want carry their expensive smart-phones with them when they exercise. I'm also always forgetting it, don't charge it, etc. These new devices are so small you can basically wear it all day and forget about it.
- hatberet 14y ago
  I'm not sure about Androids, but iPhones do not have a barometric pressure sensor, so the only way to obtain altitude data would be through GPS. For applications like this, where you don't need real-time altitude information, that low-power GPS paper from Microsoft could come in handy, but it's cheaper and lower power to just use a pressure transducer in the first place.
senthilnayagam 14y ago
I like fitbit, it is inspiring me daily to go for my workout. Minute by minute data can be very useful for custom visualization. I have found a few bugs in fitbit, like I change my daily targets, but weekly targets dont change. Also, fitbit free app is not available in Indian App Store. I dont want to create a US apple account to just use this app. If open data access is available, I would be inclined to but a Aria and couple of fitbit zip for my family members as well.
- heyitsnick 14y ago
  > like I change my daily targets, but weekly targets dont change. fwiw I think that is by design; the two things are independent. e.g. you may wish to set a minimum 10k steps a day, but a weekly goal of 100k (or whatever). It's a bit confusing because it sets you up defaults for both day and week where week=7xDay, but it doesn't have to be that way.
lispm 14y ago
I like the device. But without access to my data I will not buy it.
RossM 14y ago
I was planning on getting a FitBit this week, the main purchasing decision against a Nike Fuelband being open access to data. Does anyone have any alternatives they can recommend? I'm especially interested in the sleep tracking side of things.
- binarysolo 14y ago
  If you just want to track sleep, use Sleep Cycle (iPhone/Android phone app). I personally love the FitBit ecosystem (I use the app and scale religiously, and its UX is fantastic) and stopped using Sleep Cycle in favor of the tracking via FitBit.
jessedhillon 14y ago
The Bodymedia FIT, in addition to being a superior device overall(1), has an API which let's you access caloric burn and other data up to single minute detail levels. If you Google around, there are some blog posts describing how to dump data from it using standard USB serial port drivers. API webpage: https://developer.bodymedia.com/docs/read/Home https://developer.bodymedia.com/docs/read/Home 1- The Bodymedia gives you steps taken as well as accurate sleep and caloric burn readings.
- heyitsnick 14y ago
  Could you explain a bit more when Bodymedia FIT is superior? The FitBit One and Ultra tracks steps taken, floors climbed, and sleep efficiency, and tells you calories burned and distance travelled (the latter two i assume just simply calculated from the raw data collected; you can input your weight and stride length on the site).
  - b3b0p 14y ago
    I believe the Fitbit is nothing more than a step counter type of device. I could be wrong, but that's what it looks like. The Bodymedia FIT has actual sensors, it's an armband you have to wear. It monitors temperature among other things. It has been tested against, a $40,000 "portable oxygen analyzer", the gold standard for measuring calories. (source: http://www.bodymedia.com/the_science.html http://www.bodymedia.com/the_science.html) Regarding the Bodymedia FIT developer program. It looks you still need a subscription though and one still needs to upload the data from the armband to the website to get at it. Seems kind of pointless. I would want to get the data directly from the USB armband without a subscription.
  - jessedhillon 14y ago
    Suppose you did 100 squats. Which will be closer to reporting your true caloric burn -- a pedometer or a sensor directly contacting your skin and measuring your skin temperature and other variables. The latter, right? The former would probably report 100 steps taken, very slowly. The total costs are higher though. The Fitbit is a onetime cost of $100-125 where the Bodymedia is $120-150 once and $84/yr. But for me, the cost of the Fitbit is too high for the relatively inaccurate method of collection.
beaumartinez 14y ago
I feel the same with regards to my Nike FuelBand. I even asked Nike if they would provide an API[0] (they won't). Fortunately their web app keeps your NikeFuel—Nike's activity metric—for every day (in 15-minute increments) as a global JavaScript variable. I wrote a scraper[1] to get me my data. I wonder if you can do the same with Fitbit. [0] https://twitter.com/NikeFuel/status/205424488836370433 https://twitter.com/NikeFuel/status/205424488836370433 [1] https://github.com/beaumartinez/fuel https://github.com/beaumartinez/fuel
- marcusestes 14y ago
  They recently announced that Fuel data will be available by API after all: http://developer.nike.com/resources http://developer.nike.com/resources
  - beaumartinez 14y ago
    Thank you, you have made my day!
- qdot76367 14y ago
  I started a library to get data out of the fuelband, but never got very far mainly due to time: http://github.com/qdot/libfuelband http://github.com/qdot/libfuelband
- Domenic_S 14y ago
  > @beaumartinez Unfortunately, there will be no public API. I always wonder why company reps phrase it this way. "Unfortunately"? For us, yeah, it's unfortunate. But for Nike...?
heyitsnick 14y ago
I'm a huge fan of the entire Fitbit product; the hardware is a joy to use, it has great apps for desktop and Android, sync is a breeze, the web app works very well; the whole process has been excellent. However, having spent close to €300 with the company (for the scale + ultra, and recently upgraded to the Fitbit One), i was very disappointed when I found out about their data policy. Having to pay a premium 50/pa to get access to my own data (it's one of the big 'upsells' of premium) leaves a very bad taste in the mouth.,And now learning that even this doesn't give you access to your raw data is even worse. I hope Fitbit changes their approach; if anything I think they have it backwards: give me my data for free when I purchase your hardware for life, and upsell me the web app (i'd be happy to pay for it; maybe include the first year free with your hardware purchase); rather this than than give you a pretty complete web app for life for free to record everything i want, but and then charge me to read back my own data!
- tommoor 14y ago
  Agreed, also a fitbit owner and this seems like a no-brainer - they should open up this data and allow for more to be built on the platform ala runkeeper.
patrickk 14y ago
Fitbit's data policy seems to be in stark constrast to that of Zeo. I bought a Zeo Sleep Manager after reading Gwern's excellent, extremely detailed writeups about his sleep experiments: http://www.gwern.net/Zeo http://www.gwern.net/Zeo These wouldn't be possible without data exports: http://mysleep.myzeo.com/export/Export%20Data%20Help%20Sheet.pdf http://mysleep.myzeo.com/export/Export%20Data%20Help%20Sheet...
dadro 14y ago
I recently wrapped up a 6 month project that entailed extensive Fitbit integration. My client is a fortune 500 company and even with their resources it took them months to become a Fitbit partner. Fitbit also charged a ridiculous amount of money for accessing the 100K devices we were provisioning as compared to the competitors. If accessing the collector data is important I would consider looking at some of the fitbit competitors, as they have much more open APIs. Fitlinxx (http://www.fitlinxx.net/pebble-activity-monitor.htm http://www.fitlinxx.net/pebble-activity-monitor.htm) and Fitbug both come to mind.
rb2k_ 14y ago
Related: I also wrote a little script that uses the API to save data as YAML files: http://blog.marc-seeger.de/2012/09/16/backing-up-fitbit-data-using-their-api/ http://blog.marc-seeger.de/2012/09/16/backing-up-fitbit-data... It isn't as detailed (e.g. distribution over time) as the website, but it still gives me a warm and fuzzy feeling to have at least part of the data in machine-readable files :)
npsimons 14y ago
I also asked around and Fitbit seemed like the most “open” platform for collecting one’s own data. Either the OP didn't ask around very well, or they were lied to. There are plenty of other much more open solutions out there. Edit: A downvote? Really? In what way is the above not true?
- jessedhillon 14y ago
  Give us names, add links.
  - npsimons 14y ago
    Okay, names and links: I speak here only from personal experience; having used the following mostly at the gym, biking to and from work and while hiking in the Sierras: - Garmin Forerunner 405 (https://buy.garmin.com/shop/shop.do?pID=11039&ra=true https://buy.garmin.com/shop/shop.do?pID=11039&ra=true): All around general good "exercise" watch, but battery life requires a backup plan. Doesn't have Linux software, but can easily export to GPX for use in things like Google Earth and Maps. You can also combine other sensors: bicycle speed/cadence sensor (https://buy.garmin.com/shop/shop.do?pID=1266 https://buy.garmin.com/shop/shop.do?pID=1266) and a foot pod (https://buy.garmin.com/shop/shop.do?pID=15516 https://buy.garmin.com/shop/shop.do?pID=15516) for gym workouts, or just longer battery life. - Zephyr HXM (http://www.zephyr-technology.com/products/fitness http://www.zephyr-technology.com/products/fitness) combined with a smartphone (such as N900): Another good solution for basic logging; a variety of software to choose from, with different levels of openness. Just to curb some replies: I know that FitBit does other stuff, lasts longer on a charge, has less friction getting data off, etc, but the thing I'm countering is the openness claim. If you have to pay for your data, and you still don't get a copy, it's a) not your data and b) not an open platform. Some basic research would have revealed this, hence why I didn't (and won't) get a FitBit, and I recommend others avoid it too.
- mfringel 14y ago
  Your statement may be true, but it adds nothing to the discussion. You did not point out alternatives, examples, or links to same.
vannevar 14y ago
I was looking at FitBit but am really not interested in uploading any data. Is it possible to use the FitBit desktop app without permitting it to upload? Would it cache everything locally and function normally or does it require access to be usable at all?
- vannevar 14y ago
  Ah, found the answer in the product manual: "The Tracker will upload your data every 15 minutes provided you are within range of any plugged in base station (about 15 feet for direct line of sight), the computer is on and not in sleep or hibernate mode, the software is installed and running, and you have an active internet connection." This implies to me that the data isn't cached locally outside of the Tracker itself, if an internet connection is not available.
Too 14y ago
Because then you would be able to see how inaccurate the sensors really are?
mcormier 14y ago
Shameless plug. I've started a project to track any data that I want and stop relying on all of these free sites because of this loss of data issue. The source code is here: https://github.com/mcormier/tallyman https://github.com/mcormier/tallyman and you can see it in action here: http://stats.preenandprune.com http://stats.preenandprune.com I haven't tackled generated graphs from the SQL data yet but plan to eventually. The one graph on the page was generated with Apple's Numbers. Since it's annual data I don't plan on updating it more than once a year.
- feniv 14y ago
  I've also been working on something very similar (general data tracking with an awesome API) for a couple of months now. It's not quiet ready yet, but I would love to have your input on it once it's online! (P.S. My project is also heavily inspired by the features and shortcomings of Daytum (namely the lack of an API), which is frankly the best self tracking tool available at the moment.)
asher_ 14y ago
This is a rather annoying trend with devices in this class. I own a half dozen otherwise excellent devices like this, each of which require a SaaS subscription to make full use of, any most of which deny me access to the raw data. I have to think that the market for people wanting the raw data is fairly small in comparison to the people who just want a simple, single-device web service, but I think this is disappointing. The real value in the data is the relationships between various data sources. I really don't care how many steps I take in a particular day, but I do care if there is a relationship between that and my weight, how well I sleep, my heart rate etc. No device, or brand, does everything, nor should they. I want to be able to get my data out so I can use it in interesting ways. I have thought a lot about not buying any devices in the future that don't allow me access to the data I produce. I think it is probably a good policy, but I would be left with very little in the way of options then. Finding hacks to get data out of devices can be useful, but should we really have to crack each new gadget just to use the data? I think its a pretty sad state of affairs.
wastedbrains 14y ago
keep emailing and asking on the message boards. Fitbit is a bit slow to get back to people but normally pretty good about it. I built ruby apis to access fitbit before they had an API and a android app prior to them releasing their own. I was in email contact with them the whole time and they were willing to help me out and give me beta access to the api before it was public. I love fitbit, and hope they allow users to at least download the detailed personal reports.
cpenner461 14y ago
Just got a Jawbone UP, and I can login to my account with them and download a CSV with my data. It's just a daily summary though (i.e. not the full hour by hour or whatever interval it's recording at).
durga 14y ago
try FitFrnd: We'll soon add a button to allow you to download ALL your data in csv/excel format. https://itunes.apple.com/us/app/fitfrnd-best-weight-loss-social/id522850347?mt=8 https://itunes.apple.com/us/app/fitfrnd-best-weight-loss-soc...
hiroprot 14y ago
Slightly OT, I used to obsess about the accuracy and actual data that these devices provide (had a BodyMedia FIT and now a FitBit), until I realized that the data itself isn't what I'm most interested in. IMHO the main draw for these devices is their ability to motivate me to be more active, sleep better, etc. Historical data is interesting, but not crucial to that function (although I'd like to see FitBit make the data available). Far more important are real-time feedback mechanisms, such as notices that I achieved a goal, motivating messages ("you're almost there"), etc.
ivankirigin 14y ago
I was just looking into something similar for Jawbone Up. There is an unofficial API http://eric-blue.com/2011/11/28/jawbone-up-api-discovery/ http://eric-blue.com/2011/11/28/jawbone-up-api-discovery/
pringles 14y ago
https://groups.google.com/forum/#!searchin/fitbit-api/Intraday/fitbit-api/kV0XOcuARtU/PdL19QPxPtEJ https://groups.google.com/forum/#!searchin/fitbit-api/Intrad...
qdot76367 14y ago
I run http://openyou.org http://openyou.org, a site dedicated to reversing as many medical devices as we (or, well, at the moment, mostly I) can get our (my) hands on. So far, we've put out Emokit (emotiv epoc headset), libfitbit, libomron, and have started libfuelband and libsensewear. The main problem is just time and lack of people resources. We need more people aware that USB/Bluetooth reversing, while somewhat complex, is not all that inherently difficult. 2 or 3 people working together can easily open up a device and protocol format with a minimal amount of [insert language that supports libusb or hidapi] code. However, not many people are interested in learning that, as they just wanna get the data and run. Understandable, but that means you're waiting on one or two people to finish something that they may lose interest in. I've been contemplating trying to teach a usb/bt reversing workshop for years now, I should really get around to trying that soon.
- vitovito 14y ago
  There's not a lot of help out there for people who are willing to put the time and effort into protocol reversing but don't know where to start! EA put out a fitness video game with a wireless heart rate monitor that showed up as a normal HID device via its USB dongle. On sale it was like $40, making it a very inexpensive way to log heart rate wirelessly. But no-one had reversed the protocol, and I couldn't find any online resources to teach me how to do it myself. I'm pretty sure I threw it out, and the product is discontinued now, but I would encourage you to at least put training materials together, if not a full workshop.
  - qdot76367 14y ago
    Yeah, I will admit that information does end up being the main problem. I've been doing this long enough that it's pretty much second nature these days, and I just find myself going "but what is there to teach?". I've taught workshops on other subjects before, and usually once I get going to writing cirriculum that question answers itself in spades, though. Just need the motivation. I'm finishing up another personal project right now, then I may start on this. I suppose I don't really have much room to complain when I've already set out a solution that just needs to be done. :)
unreal37 14y ago
So this may be a counter argument to the "Hacker" culture, but here goes. Hopefully this can spark an interesting discussion. I own a fitbit, and have loved it. I don't regret paying the $99 for it. I log into the web site, and see my daily totals, and the little graph showing when my activity spikes were. At the end of a grueling day walking around New York City, I can tell my wife that we took 20,000 steps and that's a bit interesting. It's useful and interesting - only to a point. So after reading this post, my main question is "Why?" Why do you need minute-by-minute access to pedometer data? What use is it, really? The OP says, basically, "out of curiosity". OK, so hack the thing. There's a number of links for intercepting that data during the sync process. But can you fault fitbit for not providing data that noone (not even the company itself) needs? What data can the fitbit give you on a minute-by-minute basis that is remotely useful or interesting? It's just a pedometer. At the end of the day, it tells you you took 10,000 steps or whatever. It's also interesting that you walked 50 miles this month, or have walked up steps to the level of a helicopter flies. Or how many miles you've walked this year. But minute-by-minute?
- ISL 14y ago
  Spectral analysis. http://en.wikipedia.org/wiki/Spectral_density_estimation http://en.wikipedia.org/wiki/Spectral_density_estimation Without frequent sampling, you can't get access to high frequency information. http://en.wikipedia.org/wiki/Nyquist_frequency http://en.wikipedia.org/wiki/Nyquist_frequency An example (not for regular sampling, but I hope it gets the point across): if a pedometer records the time of every footfall through the day, you can discern whether your pace is faster on Tuesdays than Sundays, even if you take the same number of steps. If those data are aggregated into steps/hour, you'll never see it. The lack of open access to the data acquired by a FitBit (I'd been considering buying one) is a certain dealbreaker.
- ejain 14y ago
  The total number of steps in neat, but I'm more interested in seeing how often and how long I move (or don't move). I consider 20,000 steps followed by 6 hours of non-stop slouching in front of the computer less ideal than just 10,000 steps but never sitting for more than an hour at a time. Fitbit doesn't provide this analysis, but I can do it myself with access to minute-by-minute data.
bigfrakkinghero 14y ago
I purchased a FitBit based on it's reputation without realizing that I would have to pay to access the data that I'm collecting... deal breaker! As others have said, I understand charging for an ongoing service like analyzing the data and using their site, but not for just accessing the raw data from the device. For me, this is the first time that not owning my own data has really put me off. I'll be returning my FitBit.
ejain 14y ago
FWIW, I asked for (and got) access to Fitbit's "intraday" API without much trouble (or extra cost).
gregcohn 14y ago
Useful roundup from TC: http://techcrunch.com/2013/01/02/best-health-apps/ http://techcrunch.com/2013/01/02/best-health-apps/

211 points by KC8ZKF 14y ago

My pet peeve: I wanted to know what fitbit is, but the site redirects me to fitbit.com/de and I closed that site. How can they be open about data, if they have obviously no clue about open standards, like http accept headers? Worse, there's no way to visit the (official?) fitbit.com site even after that idiotic redirect. I can't access an english version, period. Stupid.

ig1 14y ago
There's a logical flaw in your reasoning "doesn't http accept headers" doesn't imply anything about their understanding of open standards. While it's perfectly understandable that you might be upset that their site for Germany doesn't offer english language as an option, that's clearly a completely separate issue from how they handle user data.
- darklajid 14y ago
  You're correct. I let my anger about ignoring my explicit preferences, which I expressed in a standard way, lead me to believe that the company doesn't care. Obviously that's annoyance speaking and speculating based on me being tired with the constant 'Hey, I know better what you want' attitude (Looking at you, Google/Blogger) - I don't even know a single thing about their product (which is why I went to the site in the first place). Sorry about that. I stand by my point about this sort of redirection being a telltale sign of a flawed web site and lack of respect for user preferences. What else they can or cannot do, I have no clue about and cannot judge.
  - ig1 14y ago
    Given that you're in Germany and they don't have an english version of their Germany website available their two options would be either to send you to the German language website for Germany (which they do) or alternatively send you to the US or UK website which are in english. I'm not sure it's obvious that sending someone in Germany to the UK/US website (where presumably they'll be unable to buy the product) just because their browser is set to English is the better solution. While for sites like Blogger language is obviously more important than country, for companies which are country-localized (i.e they treat different countries differently for shipping, taxes, legal, operations, etc.) I would guess that it makes more sense to send you to local country version. Imagine you were using say a dating site or a takeaway site, you would find it equally frustrating if you were routed away from a local language site to a US specific site just because that's what your language preferences were set to.
    - darklajid 14y ago
      I'm not sure why you're conflating a) content and localization b) the region I'd like the product to ship to, if at all Why is the site different for different countries, ignoring 'translation'? I haven't thought it through, maybe, but I cannot come up with any decent reason for a 'German' site that isn't just the 'US' site in a different language. In that case, please (dear website) listen to what I'm asking for. If I go to fitbit.com I expect to get the very same thing someone in the US receives. I'd like to talk about the very same site. I don't want someone to redirect me to a localized thing. And certainly not without giving me the opportunity to say 'Yeah, no. That was stupid. I really wanted the original version, silly'. Same thing: If I go to www.google.com, I want to end up at www.google.com, not www.google.de. If I visit a random post on Blogger, chances are everything content is in English. Except for the 'helpful' Blogger toolbar and whatnot, that are coming up in German, because hey that's where we figured out you're coming from. Lived in Israel for a year, got a Hebrew toolbar, google.il (and I'd like to know what fitbits would've done there). German vs. English is one thing: I can read both, I just explicitly (url, domain, accept headers) ask for the latter. English vs. Hebrew is another: I cannot read the latter, even if I happen to be - yay for geolocation - in the one state that represents the Hebrew language. Imagine you were using say a dating site or a takeaway site, you would find it equally frustrating if you were routed away from a local language site to a US specific site Right. Don't send me anywhere if I navigate to example.com, even if I ask for de_DE. Offer a translated version, if you can. Otherwise drop a small (German?) link on your .com, saying 'We noticed you explicitly ask for German content. We got a country specific site right here -> example.de' A dating site would allow me to register and state my country of origin or interest (which might be Germany, even if I live in Tel Aviv at that time). A takeaway site is really a weird example. www.pizza.de is available in German only for all I can tell and won't redirect me to a random US site because I ask the server to please return en_US or en localized content, _if possible_. So, for me this whole 'automagic-we-know-it-best' translation/redirection thing is broken by design. It was a constant hassle in the past and just seems to catch on. Which is why I'm pointing it out when I can. I'm sorry for the thread-jacking. Thanks for the exchange so far.

gregcohn 14y ago

Good post. Are are any of the similar devices more open, in terms of both data accessibility and format?

jrockway 14y ago
If a FitBit is just a pedometer, I have an Omron HJ-720IT that syncs with my Linux box (uploading a week's worth of hourly step counts). It's $30 on Amazon.
- thisone 14y ago
  The cheaper one is a pedometer, the expensive one tries to do things like track your sleeping patterns and hooks into their own make of scale.

mavlee 14y ago

I actually ran into the same problem recently, when trying to get my minute by minute data. I even emailed their support team, but no luck. I wonder if there's an easy way to scrape their flash charts on the dashboard...

MichaelApproved 14y ago
Without using the website, I suspect you can use something like Fiddler to see what HTTP requests are being made. Chances are the charts are making an XML or JSON request to their servers. If you're lucky, the URL requests will be in an easy to reverse engineer format so you could easily dump all the data you need by adjusting the URL.
- bloaf 14y ago
  Why not try to intercept the data as it comes through the USB port instead?
  - Thrymr 14y ago
    Or try liftbit as noted above: https://github.com/openyou/libfitbit https://github.com/openyou/libfitbit
    - simcop2387 14y ago
      libfitbit unfortunately doesn't currently work with the fitbit one, for the older devices it's apparently fine but the one changes how it gets talked to entirely. I've attempted a bit of work with getting something talking to it but it's not an easy thing to do. Their bluetooth dongle presents itself as an HID device and I think hides a lot of the details away from you if you wanted to talk to it over actual bluetooth. I'd love to get it to sync with both their site and to be able to capture the information myself to store.
fla 14y ago
The flash content must get the data from somewhere, probably in XML, certainly over HTTP. Open Wireshark and capture a session to known where it came from. Then write a program that replicates this behavior and save the data.

harold 14y ago

This project hasn't been updated in a while but might be a good starting point for a roll-your-own solution: https://github.com/loghound/Fitbit-for-Google-App-Script https://github.com/loghound/Fitbit-for-Google-App-Script

nodata 14y ago

Just wait until a large health insurance company buys Fitbit (or Withings, or Garmin) - they will have your data.

jrockway 14y ago
Just wait until a large health insurance company breaks into your house at night and installs cameras - they will have your data.
alexkus 14y ago
I have various Garmins (Forerunner 405 for running and Edge 705 for cycling). None of my data goes to Garmin as I just use Garmin Training Centre and don't upload anything to Garmin Connect.
- npsimons 14y ago
  Much as I dislike Garmin's lack of support for Linux, I have to second this. My Forerunner 405, eTrex Vista and Vista HCx all download to my computer, with the option (not forced) of uploading. I have full access to .GPX files that I can use with any number of pieces of software (including Google Earth and Maps). Also, I've been pretty happy with the Zephyr HxM hooked up to my N900, and the wife has a Zephyr as well that works well with a Samsung Galaxy III and Google tracks.
  - quantumstate 14y ago
    It case you haven't come across it I have found that garmintools and pytrainer work very well with my forerunner 305. I have never used anything else.

MichaelApproved 14y ago

"I guess it is true, if you aren’t paying for it, you are the product." But according to the author, even the $50 option doesn't let you download the data so it has nothing to do with you being the product.

vertis 14y ago
More than that, you have to pay for the device in the first place, so on that front you're absolutely the customer.
- jrockway 14y ago
  Indeed. Everyone likes to think missing features are some scam that the company that owns the project is perpetrating upon their users, but in reality they probably want to lock you in for the sake of locking you in. They don't know why, but making the data open would be extra work, would help their competitors, and might hurt their chances of being bought. They can always be more open later, but they can't ever un-open. So they're closed. Don't buy their products. That's what I did.
  - angdis 14y ago
    I agree, there are other reasons for not providing the raw data. For the vast majority of people raw-data just isn't something that is even remotely usable. Even if one is an expert with data analysis, IMHO, it will require a lot of work to process and interpret it correctly. It may even require a deep knowledge about the device and all its quirks-- we're talking about data derived from a cheap accelerometer here. In the end, it is hard to see how opening up the raw data so a few hackers can take a crack at it is advantageous to fitbit or even consumers in general.

randomchars 14y ago

If you live in the EU you can send them a data access request, which means they have to provide you with all your personally identifiable information that they keep, in this case that means pretty much everything. I hate it when companies think that they can get away with making you pay for your data.

dawson 14y ago
I have always been curious about this, from a healthcare perspective. For example, we encrypt PID information one way (the patient has their own unique data access key, which is then encrypted with their password + salt and some 'other stuff'), so we couldn't hand over any information if a data access request was made, even if we wanted to as we don't have access to that information ourselves!
- hayksaakian 14y ago
  You could hand over the encrypted data, the usefulness of such data being low however.
  - dawson 14y ago
    The issue is that the PID is all encrypted, I wouldn't know if we have a John Doe on our system, we can't see anything to verify that user exists or that they're the appropriate owner of said data. We do have an email (which is also encrypted at a system level, which in theory we could access), but verifying an access request based on an email address? I don't know. The way we solve the problem is by making all the data available to the user once they have authenticated, as long as you can login to our website, all your data is available to you, but if you made a direct data access request to us, like a FOI to our office by letter (the typical way it's done), not much I think we could do.
    - thisone 14y ago
      Your IG team will know. From my understanding, from the yearly IG briefings I used to have, you would be able to provide that requester with information on how to obtain their personal access, and point out to them where they can find the information they have requested. Information needs to be made accessible, you don't necessarily need to actually provide printouts. For example, FOI requests from news agencies often cover the same topics. This information can and often is posted publicly on the organisation's website and the FOI responses refer the requester to those links.

yarrel 14y ago

There's libfitbit - https://github.com/openyou/libfitbit https://github.com/openyou/libfitbit

rb2k_ 14y ago
The new fitbit can sync using the iphone's bluetooth 4 connection. This makes syncing a passive rather than an active thing which was the reason I bought the fitbit one even though I had the original one.

biturd 14y ago

Why do we need an actual hardware device to get to this data? Don't iPhones and Androids and others have all the internals needed to get the same data, and more, such as altitude, so you can check the incline you are against etc.

heyitsnick 14y ago
I think for lots of people, they don't want carry their expensive smart-phones with them when they exercise. I'm also always forgetting it, don't charge it, etc. These new devices are so small you can basically wear it all day and forget about it.
hatberet 14y ago
I'm not sure about Androids, but iPhones do not have a barometric pressure sensor, so the only way to obtain altitude data would be through GPS. For applications like this, where you don't need real-time altitude information, that low-power GPS paper from Microsoft could come in handy, but it's cheaper and lower power to just use a pressure transducer in the first place.

senthilnayagam 14y ago

I like fitbit, it is inspiring me daily to go for my workout. Minute by minute data can be very useful for custom visualization. I have found a few bugs in fitbit, like I change my daily targets, but weekly targets dont change. Also, fitbit free app is not available in Indian App Store. I dont want to create a US apple account to just use this app. If open data access is available, I would be inclined to but a Aria and couple of fitbit zip for my family members as well.

heyitsnick 14y ago
> like I change my daily targets, but weekly targets dont change. fwiw I think that is by design; the two things are independent. e.g. you may wish to set a minimum 10k steps a day, but a weekly goal of 100k (or whatever). It's a bit confusing because it sets you up defaults for both day and week where week=7xDay, but it doesn't have to be that way.

lispm 14y ago

I like the device. But without access to my data I will not buy it.

RossM 14y ago

I was planning on getting a FitBit this week, the main purchasing decision against a Nike Fuelband being open access to data. Does anyone have any alternatives they can recommend? I'm especially interested in the sleep tracking side of things.

binarysolo 14y ago
If you just want to track sleep, use Sleep Cycle (iPhone/Android phone app). I personally love the FitBit ecosystem (I use the app and scale religiously, and its UX is fantastic) and stopped using Sleep Cycle in favor of the tracking via FitBit.

jessedhillon 14y ago

The Bodymedia FIT, in addition to being a superior device overall(1), has an API which let's you access caloric burn and other data up to single minute detail levels. If you Google around, there are some blog posts describing how to dump data from it using standard USB serial port drivers. API webpage: https://developer.bodymedia.com/docs/read/Home https://developer.bodymedia.com/docs/read/Home 1- The Bodymedia gives you steps taken as well as accurate sleep and caloric burn readings.

heyitsnick 14y ago
Could you explain a bit more when Bodymedia FIT is superior? The FitBit One and Ultra tracks steps taken, floors climbed, and sleep efficiency, and tells you calories burned and distance travelled (the latter two i assume just simply calculated from the raw data collected; you can input your weight and stride length on the site).
- b3b0p 14y ago
  I believe the Fitbit is nothing more than a step counter type of device. I could be wrong, but that's what it looks like. The Bodymedia FIT has actual sensors, it's an armband you have to wear. It monitors temperature among other things. It has been tested against, a $40,000 "portable oxygen analyzer", the gold standard for measuring calories. (source: http://www.bodymedia.com/the_science.html http://www.bodymedia.com/the_science.html) Regarding the Bodymedia FIT developer program. It looks you still need a subscription though and one still needs to upload the data from the armband to the website to get at it. Seems kind of pointless. I would want to get the data directly from the USB armband without a subscription.
- jessedhillon 14y ago
  Suppose you did 100 squats. Which will be closer to reporting your true caloric burn -- a pedometer or a sensor directly contacting your skin and measuring your skin temperature and other variables. The latter, right? The former would probably report 100 steps taken, very slowly. The total costs are higher though. The Fitbit is a onetime cost of $100-125 where the Bodymedia is $120-150 once and $84/yr. But for me, the cost of the Fitbit is too high for the relatively inaccurate method of collection.

beaumartinez 14y ago

I feel the same with regards to my Nike FuelBand. I even asked Nike if they would provide an API[0] (they won't). Fortunately their web app keeps your NikeFuel—Nike's activity metric—for every day (in 15-minute increments) as a global JavaScript variable. I wrote a scraper[1] to get me my data. I wonder if you can do the same with Fitbit. [0] https://twitter.com/NikeFuel/status/205424488836370433 https://twitter.com/NikeFuel/status/205424488836370433 [1] https://github.com/beaumartinez/fuel https://github.com/beaumartinez/fuel

marcusestes 14y ago
They recently announced that Fuel data will be available by API after all: http://developer.nike.com/resources http://developer.nike.com/resources
- beaumartinez 14y ago
  Thank you, you have made my day!
qdot76367 14y ago
I started a library to get data out of the fuelband, but never got very far mainly due to time: http://github.com/qdot/libfuelband http://github.com/qdot/libfuelband
Domenic_S 14y ago
> @beaumartinez Unfortunately, there will be no public API. I always wonder why company reps phrase it this way. "Unfortunately"? For us, yeah, it's unfortunate. But for Nike...?

heyitsnick 14y ago

I'm a huge fan of the entire Fitbit product; the hardware is a joy to use, it has great apps for desktop and Android, sync is a breeze, the web app works very well; the whole process has been excellent. However, having spent close to €300 with the company (for the scale + ultra, and recently upgraded to the Fitbit One), i was very disappointed when I found out about their data policy. Having to pay a premium 50/pa to get access to my own data (it's one of the big 'upsells' of premium) leaves a very bad taste in the mouth.,And now learning that even this doesn't give you access to your raw data is even worse. I hope Fitbit changes their approach; if anything I think they have it backwards: give me my data for free when I purchase your hardware for life, and upsell me the web app (i'd be happy to pay for it; maybe include the first year free with your hardware purchase); rather this than than give you a pretty complete web app for life for free to record everything i want, but and then charge me to read back my own data!

tommoor 14y ago
Agreed, also a fitbit owner and this seems like a no-brainer - they should open up this data and allow for more to be built on the platform ala runkeeper.

patrickk 14y ago

Fitbit's data policy seems to be in stark constrast to that of Zeo. I bought a Zeo Sleep Manager after reading Gwern's excellent, extremely detailed writeups about his sleep experiments: http://www.gwern.net/Zeo http://www.gwern.net/Zeo These wouldn't be possible without data exports: http://mysleep.myzeo.com/export/Export%20Data%20Help%20Sheet.pdf http://mysleep.myzeo.com/export/Export%20Data%20Help%20Sheet...

dadro 14y ago

I recently wrapped up a 6 month project that entailed extensive Fitbit integration. My client is a fortune 500 company and even with their resources it took them months to become a Fitbit partner. Fitbit also charged a ridiculous amount of money for accessing the 100K devices we were provisioning as compared to the competitors. If accessing the collector data is important I would consider looking at some of the fitbit competitors, as they have much more open APIs. Fitlinxx (http://www.fitlinxx.net/pebble-activity-monitor.htm http://www.fitlinxx.net/pebble-activity-monitor.htm) and Fitbug both come to mind.

rb2k_ 14y ago

Related: I also wrote a little script that uses the API to save data as YAML files: http://blog.marc-seeger.de/2012/09/16/backing-up-fitbit-data-using-their-api/ http://blog.marc-seeger.de/2012/09/16/backing-up-fitbit-data... It isn't as detailed (e.g. distribution over time) as the website, but it still gives me a warm and fuzzy feeling to have at least part of the data in machine-readable files :)

npsimons 14y ago

I also asked around and Fitbit seemed like the most “open” platform for collecting one’s own data. Either the OP didn't ask around very well, or they were lied to. There are plenty of other much more open solutions out there. Edit: A downvote? Really? In what way is the above not true?

jessedhillon 14y ago
Give us names, add links.
- npsimons 14y ago
  Okay, names and links: I speak here only from personal experience; having used the following mostly at the gym, biking to and from work and while hiking in the Sierras: - Garmin Forerunner 405 (https://buy.garmin.com/shop/shop.do?pID=11039&ra=true https://buy.garmin.com/shop/shop.do?pID=11039&ra=true): All around general good "exercise" watch, but battery life requires a backup plan. Doesn't have Linux software, but can easily export to GPX for use in things like Google Earth and Maps. You can also combine other sensors: bicycle speed/cadence sensor (https://buy.garmin.com/shop/shop.do?pID=1266 https://buy.garmin.com/shop/shop.do?pID=1266) and a foot pod (https://buy.garmin.com/shop/shop.do?pID=15516 https://buy.garmin.com/shop/shop.do?pID=15516) for gym workouts, or just longer battery life. - Zephyr HXM (http://www.zephyr-technology.com/products/fitness http://www.zephyr-technology.com/products/fitness) combined with a smartphone (such as N900): Another good solution for basic logging; a variety of software to choose from, with different levels of openness. Just to curb some replies: I know that FitBit does other stuff, lasts longer on a charge, has less friction getting data off, etc, but the thing I'm countering is the openness claim. If you have to pay for your data, and you still don't get a copy, it's a) not your data and b) not an open platform. Some basic research would have revealed this, hence why I didn't (and won't) get a FitBit, and I recommend others avoid it too.
mfringel 14y ago
Your statement may be true, but it adds nothing to the discussion. You did not point out alternatives, examples, or links to same.

vannevar 14y ago

I was looking at FitBit but am really not interested in uploading any data. Is it possible to use the FitBit desktop app without permitting it to upload? Would it cache everything locally and function normally or does it require access to be usable at all?

vannevar 14y ago
Ah, found the answer in the product manual: "The Tracker will upload your data every 15 minutes provided you are within range of any plugged in base station (about 15 feet for direct line of sight), the computer is on and not in sleep or hibernate mode, the software is installed and running, and you have an active internet connection." This implies to me that the data isn't cached locally outside of the Tracker itself, if an internet connection is not available.

Too 14y ago

Because then you would be able to see how inaccurate the sensors really are?

mcormier 14y ago

Shameless plug. I've started a project to track any data that I want and stop relying on all of these free sites because of this loss of data issue. The source code is here: https://github.com/mcormier/tallyman https://github.com/mcormier/tallyman and you can see it in action here: http://stats.preenandprune.com http://stats.preenandprune.com I haven't tackled generated graphs from the SQL data yet but plan to eventually. The one graph on the page was generated with Apple's Numbers. Since it's annual data I don't plan on updating it more than once a year.

feniv 14y ago
I've also been working on something very similar (general data tracking with an awesome API) for a couple of months now. It's not quiet ready yet, but I would love to have your input on it once it's online! (P.S. My project is also heavily inspired by the features and shortcomings of Daytum (namely the lack of an API), which is frankly the best self tracking tool available at the moment.)

asher_ 14y ago

This is a rather annoying trend with devices in this class. I own a half dozen otherwise excellent devices like this, each of which require a SaaS subscription to make full use of, any most of which deny me access to the raw data. I have to think that the market for people wanting the raw data is fairly small in comparison to the people who just want a simple, single-device web service, but I think this is disappointing. The real value in the data is the relationships between various data sources. I really don't care how many steps I take in a particular day, but I do care if there is a relationship between that and my weight, how well I sleep, my heart rate etc. No device, or brand, does everything, nor should they. I want to be able to get my data out so I can use it in interesting ways. I have thought a lot about not buying any devices in the future that don't allow me access to the data I produce. I think it is probably a good policy, but I would be left with very little in the way of options then. Finding hacks to get data out of devices can be useful, but should we really have to crack each new gadget just to use the data? I think its a pretty sad state of affairs.

wastedbrains 14y ago

keep emailing and asking on the message boards. Fitbit is a bit slow to get back to people but normally pretty good about it. I built ruby apis to access fitbit before they had an API and a android app prior to them releasing their own. I was in email contact with them the whole time and they were willing to help me out and give me beta access to the api before it was public. I love fitbit, and hope they allow users to at least download the detailed personal reports.

cpenner461 14y ago

Just got a Jawbone UP, and I can login to my account with them and download a CSV with my data. It's just a daily summary though (i.e. not the full hour by hour or whatever interval it's recording at).

durga 14y ago

try FitFrnd: We'll soon add a button to allow you to download ALL your data in csv/excel format. https://itunes.apple.com/us/app/fitfrnd-best-weight-loss-social/id522850347?mt=8 https://itunes.apple.com/us/app/fitfrnd-best-weight-loss-soc...

hiroprot 14y ago

Slightly OT, I used to obsess about the accuracy and actual data that these devices provide (had a BodyMedia FIT and now a FitBit), until I realized that the data itself isn't what I'm most interested in. IMHO the main draw for these devices is their ability to motivate me to be more active, sleep better, etc. Historical data is interesting, but not crucial to that function (although I'd like to see FitBit make the data available). Far more important are real-time feedback mechanisms, such as notices that I achieved a goal, motivating messages ("you're almost there"), etc.

ivankirigin 14y ago

I was just looking into something similar for Jawbone Up. There is an unofficial API http://eric-blue.com/2011/11/28/jawbone-up-api-discovery/ http://eric-blue.com/2011/11/28/jawbone-up-api-discovery/

pringles 14y ago

https://groups.google.com/forum/#!searchin/fitbit-api/Intraday/fitbit-api/kV0XOcuARtU/PdL19QPxPtEJ https://groups.google.com/forum/#!searchin/fitbit-api/Intrad...

qdot76367 14y ago

I run http://openyou.org http://openyou.org, a site dedicated to reversing as many medical devices as we (or, well, at the moment, mostly I) can get our (my) hands on. So far, we've put out Emokit (emotiv epoc headset), libfitbit, libomron, and have started libfuelband and libsensewear. The main problem is just time and lack of people resources. We need more people aware that USB/Bluetooth reversing, while somewhat complex, is not all that inherently difficult. 2 or 3 people working together can easily open up a device and protocol format with a minimal amount of [insert language that supports libusb or hidapi] code. However, not many people are interested in learning that, as they just wanna get the data and run. Understandable, but that means you're waiting on one or two people to finish something that they may lose interest in. I've been contemplating trying to teach a usb/bt reversing workshop for years now, I should really get around to trying that soon.

vitovito 14y ago
There's not a lot of help out there for people who are willing to put the time and effort into protocol reversing but don't know where to start! EA put out a fitness video game with a wireless heart rate monitor that showed up as a normal HID device via its USB dongle. On sale it was like $40, making it a very inexpensive way to log heart rate wirelessly. But no-one had reversed the protocol, and I couldn't find any online resources to teach me how to do it myself. I'm pretty sure I threw it out, and the product is discontinued now, but I would encourage you to at least put training materials together, if not a full workshop.
- qdot76367 14y ago
  Yeah, I will admit that information does end up being the main problem. I've been doing this long enough that it's pretty much second nature these days, and I just find myself going "but what is there to teach?". I've taught workshops on other subjects before, and usually once I get going to writing cirriculum that question answers itself in spades, though. Just need the motivation. I'm finishing up another personal project right now, then I may start on this. I suppose I don't really have much room to complain when I've already set out a solution that just needs to be done. :)

unreal37 14y ago

So this may be a counter argument to the "Hacker" culture, but here goes. Hopefully this can spark an interesting discussion. I own a fitbit, and have loved it. I don't regret paying the $99 for it. I log into the web site, and see my daily totals, and the little graph showing when my activity spikes were. At the end of a grueling day walking around New York City, I can tell my wife that we took 20,000 steps and that's a bit interesting. It's useful and interesting - only to a point. So after reading this post, my main question is "Why?" Why do you need minute-by-minute access to pedometer data? What use is it, really? The OP says, basically, "out of curiosity". OK, so hack the thing. There's a number of links for intercepting that data during the sync process. But can you fault fitbit for not providing data that noone (not even the company itself) needs? What data can the fitbit give you on a minute-by-minute basis that is remotely useful or interesting? It's just a pedometer. At the end of the day, it tells you you took 10,000 steps or whatever. It's also interesting that you walked 50 miles this month, or have walked up steps to the level of a helicopter flies. Or how many miles you've walked this year. But minute-by-minute?

ISL 14y ago
Spectral analysis. http://en.wikipedia.org/wiki/Spectral_density_estimation http://en.wikipedia.org/wiki/Spectral_density_estimation Without frequent sampling, you can't get access to high frequency information. http://en.wikipedia.org/wiki/Nyquist_frequency http://en.wikipedia.org/wiki/Nyquist_frequency An example (not for regular sampling, but I hope it gets the point across): if a pedometer records the time of every footfall through the day, you can discern whether your pace is faster on Tuesdays than Sundays, even if you take the same number of steps. If those data are aggregated into steps/hour, you'll never see it. The lack of open access to the data acquired by a FitBit (I'd been considering buying one) is a certain dealbreaker.
ejain 14y ago
The total number of steps in neat, but I'm more interested in seeing how often and how long I move (or don't move). I consider 20,000 steps followed by 6 hours of non-stop slouching in front of the computer less ideal than just 10,000 steps but never sitting for more than an hour at a time. Fitbit doesn't provide this analysis, but I can do it myself with access to minute-by-minute data.

bigfrakkinghero 14y ago

I purchased a FitBit based on it's reputation without realizing that I would have to pay to access the data that I'm collecting... deal breaker! As others have said, I understand charging for an ongoing service like analyzing the data and using their site, but not for just accessing the raw data from the device. For me, this is the first time that not owning my own data has really put me off. I'll be returning my FitBit.

ejain 14y ago

FWIW, I asked for (and got) access to Fitbit's "intraday" API without much trouble (or extra cost).

gregcohn 14y ago

Useful roundup from TC: http://techcrunch.com/2013/01/02/best-health-apps/ http://techcrunch.com/2013/01/02/best-health-apps/