QS Show & Tell III

Gary Wolf

The next Quantified Self Show&Tell is December 9, 2008, hosted by IDEO-San Francisco . As at the last two events, the program will be simple. Sign up to talk about your self-tracking/personal data/life-logging project either via email to me or Kevin or by talking to us before the meeting. Presentations are brief and informal.

So much good conversation got started at the last meeting that we decided to give some time for people to talk before the program starts. Our hosts at IDEO have offered their space starting at 6 p.m. The program will get going promptly at 7.

All the details of the Show&Tell are on our Quantified Self MeetUp page. Please RSVP so we know how many to expect.

Ryan Grant - TIVO for Life

Gary Wolf

Here is a great talk by Ryan Grant from the last QS Show&Tell. Things got especially interesting when Ryan talked started talking about how the device he is making would allow you to capture, in sound, stills, and video, moments of your life that had already passed.

 Below are a few excerpts from the audio transcript to whet your interest.

KK: "Quiet please!"

RG: "Hi my name is Ryan Grant, I'm the founder of Metascopic, Incorporated. We make a memory assistant in the form of a camera you can wear. It takes tens of thousands of pictures during the day and records audio as well. I"m really excited we can get it down to a product about this size. It very wearable. This has not been done before, and I don't know why...."

QUESTION FROM AUDIENCE: "Tell us a little more abou the specs. What is this? Is this a still camera, is it video? I'm not really sure what it is yet, can you describe what it hopes to be?"

RG: "The first thing I can tell you is that you are going to get  tens of thousands of pictures a day."

QUESTION FROM AUDIENCE: "Still pictures?"

RG: "Still pictures."

QUESTION FROM AUDIENCE: "What resolution?"

RG: "XGA resolution. 1024 by 768."

QUESTION FROM AUDIENCE: "What about the viewpoint?"

RG: The viewpoint is going to be very wide angle. For some reason nobody is doing this in existing semi-wearables.

QUESTION FROM AUDIENCE: How often per minute does that translate to?

RG: That translates to a picture every 2 to 5 seconds.

Now, using a buffer, you could tap the device and capture the moment that had just passed - a kind of TIVO for life.



 

QS_081023_02_Ryan_Grant from Paul Lundahl on Vimeo.

Self-Tracking Through SMS

Gary Wolf

Just a quick follow up to the last post about Tweet What You Eat, inspired again by Flowing Data and by a telling anecdote from a recent health conference, where I concluded that ubiquitous self-tracking is coming, but perhaps not from the direction expected by many health professionals.

At the conference I met the CEO of health informatics company who had a seemingly clever idea about how to collect patient data, track compliance with medical recommendations (prescriptions, glucose monitoring, etc.), and provide reminders to patients and reports to their doctors. His name is Kent Dicks, and the company is called MedApps. Dicks is smart and well informed, and is working hard to make ubiquitous tracking work.

MedApps uses early generation wireless, where the connectivity costs are cheap and the bandwidth meager. The business idea was the create a general device, basically a wireless transmitter, that would transfer data over the cellular networks, and then arm this devices with a bunch of different customized dongles to connect to all different types of measurement tools: scale, glucose monitor, pedometer, cardiac monitor, etc. Owning a general, proprietary solution to an admittedly messy problem justifies a multi-hundred dollar price tag. After all, this is meant to be a medical system, approved by the FDA and paid for by the great mother of health care business plans: insurance reimbursements. Patients would comply with self-tracking because they wanted to cooperate with the regimen given to them by their doctors. If added motivation was needed, they could be influenced by insurance discounts, or even paycheck bonuses or penalties in the context of corporate "wellness" programs.

I long for a general solution to the data collection problem, but this approach strikes me as somewhat  wrong headed. It is serving the labyrinthine business system, rather than the more straightforward and obvious needs of people. (People who are not necessarily "patients.") Meanwhile, companies like Zume Life are just having people read their data into a digital recorder and transcribing the reports, skilled users like Nathan Yau are writing Twitter bots, and small scale entrepreneurs like Alex Rossi are creating Web based services like Tweet What You Eat that collect reports via SMS. Data is beginning to flow through and around the highly disorganized and loosely connected networks that already exist, and as this flow increases I wonder if the more fully engineered and FDA approved systems will simply become irrelevant. And the bigger theme is also relevant. Either self-tracking will be understood as a good thing, and many people will want to do it, or it will not be so understood, and all the inducements in the world will probably not be enough to motivate them.

To get a sense of how low the threshold is for programmers to build a simple system to gather personal data, take a look at Nathan's recent post on Flowing Data about how to make your own Twitter bot. Also read the comments. When I did, I thought: "Pioneering users can make popular programs... This problem is on the way to being solved... "

Alex Rossi Shows Tweet What You Eat

Gary Wolf

This post makes me happy! One of the most fun things about QS so far has been the sense of optimism and possibility emanating from the frontiers of self-tracking. There is something so obvious about applying basic methods of rational data gathering and analysis to daily life that each little experiment, however simple, hints at bigger themes.

At the last QS Show&Tell, Alex Rossi showed his Twitter apps Tweet What You Eat and Tweet What You Spend. Since then, Nathan Yau of Flowing Data has posted about this attempt to track his eating through Twitter. Nathan wrote a little bot to collect Twitter messages about what he's eating and how much he weighs and stick them in a database, which he can then use to chart his progress. He asked on his blog whether, if he made this public, people would be interested in using it.

Alex Rossi's experience suggests that, yes, people would be interested. Tweet What You Eat and Tweet What You Spend are free apps Rossi wrote that do similar work: take SMS messages and post them to a database. Rossi has added some good tricks, such as crowd-sourcing the calorie count, so that suggested values are quickly available. But what I enjoyed most of Alex's presentation was how clearly he outlined the power of this simple tool. I had just come from the Health 2.0 conference, where there was discussion of all kinds of complex mechanisms for gathering and presenting patient data. Devices, networks, payment systems, regulations - who was going to solve the puzzle? And then down to the QS Show&Tell, where one intelligent person, using a pared down protocol and an extremely simple social networking platform, hinted at a solution that is just around the corner, and that can't be seen from the perspective of "health care."

Anyway, here's the video. My favorite quote: "I noticed people would debit exercise from their food diary. I was like, I didn't even know I supported negative values!"





QS_081023_06_Alex_Rossi from Paul Lundahl on Vimeo.

Can You Eat Yourself Smarter?

Gary Wolf

Here is a great presentation by Tim Lundeen from the recent QS Show&Tell. Tim is running some interesting self-experiments on diet and cognition.


QS_081023_03_Tim_Lundeen from Paul Lundahl on Vimeo.

Diet and cognition is a topic of such obvious interest that it regularly breaks through into the popular press and the science blogs. For instance, eating blueberries and walnuts make rats smarter.

Tim was specifically inspired by the many posts Seth Roberts has made about the benefits of omega-3 fatty acids. Tim uses a simple test of cognitive function as his dependent variable: he gives himself 100 very simple math problems and records the time it takes to complete them. Here is one of the graphs Tim made in his self-study. The y-axis is the time it take to complete his 100 problems. On about day 80, he upped his dose of DHA from fish oil.

QS Show & Tell II

Gary Wolf

Last Thursday's Quantified Self Show & Tell saw some great presentations, with great questions and discussion - or rather the beginning of what could have been much longer discussions that we cut off every time out of enthusiasm for the next person's show & tell. Average presentation time was a little under ten minutes, average discussion time was also under ten minutes, which allowed us to hear from 8 people in two hours. We cleared the room at 10 p.m. in order not to further abuse the extremely gracious hospitality of our hosts at The Institute for the Future, but as I was helping carry video equipment to the parking lot I noticed that the conversation has moved itself outside and didn't show any sign of diminishing.



Below is a list of who presented, along with a single sentence about the topic. Fortunately, Paul Lundahl of eMotion studios, a member of the QS Show&Tell gang, was at the meeting with his digital video setup, and over the next few weeks we will be publishing short videos of some of the presentations.

Thanks to all who presented. We're going to try to make the QS Show&Tell a monthly event, using different interesting venues. (If you'd like to hear about them, you can sign up here: Quantified Self Show&Tell Meetup.)

Here's a brief summary of what happened:

Faren talked about her quest for a self-cure and tracking more than 40 biometric indicators.

Brandon from A&D Weighing gave us a demo of a collection of  biometric devices that will automatically upload data to a proprietary web site for analysis.

Steve gave a short presentation raising the possibility that Bayesian analysis could help self-quantifiers who have trouble keeping testing conditions under rigid control (i.e., all of us).

Break here to say that enthusiasm for Bayesian analysis has broken out at both Quantified Self Show&Tells, raising the happiness of some and leaving others baffled. This will be a topic for future posts, but here are two quick links. The first is a wikipedia entry with links to many subtopics. The second is a beautiful, lengthy conceptual tutorial that is pretty accessible.
Wikipedia on Bayesian stuff.
Eliezer S. Yudkowsky, "An Intuitive Explanation of Bayes Theorem."

Ryan Grant gave an impassioned and inspiring talk about life-logging with an always on wearable camera he is developing.

Tim Lundeen talked about tracking cognitive function using a standard, easy to implement test and correlating changes in diet with changes in cognition.

Luke showed a life logging tool he created using Filemaker, which he uses to correlate life events with other types of time lines, such as world events, and life events of other family members and friends.

Joe Betts-Lacroix described his dream architecture for a self tracking system.

Alex Ressi gave a wonderful demo of his sites TweetWhatYouEat and TweetWhatYouSpent that use SMS/Twitter to post personal tracking data.

Look for video of some of these presentations in the next few weeks, along with more details!

QS Wiki Invitation

Gary Wolf

This is a quick post to invite QS readers to contribute to the Quantified Self Wiki. The Wiki address is:

http://quantifiedself.wik.is/

You must register but registration is open and you can begin contributing as soon as you create a user name - no need to wait for a confirmation email or any other bureaucracy at the moment.

Much of the wiki is intuitive, but editing side panels and organizational-level structure can require some tricks. If you have questions, feel free to email research librarian Camille Cloutier who has read the docs!

Second Quantified Self Show-n-Tell

Kevin Kelly

The second official Quantified Self Show and Tell will take place this Thursday evening, Oct 23. 

Our first meeting last month exceeded our expectations, both in the number of people who came and the sophistication of the self-tracking projects that were shared and discussed. It was a real blast. Almost 30 folks showed up. So we decided to do it again. Since the last meeting maxed out the capacity of the location in my studio, with people sitting on the stairs, we are holding this month's Show&Tell, appropriately, at The Institute for the Future, in Palo Alto, where there is room enough for all. (The Institute for the Future is a general purpose consultancy built around future studies.)

The format will be simple. We will have some extremely brief introductions, list some of the areas of interest, and move on to the Show&Tell. If you are self-tracking in any way - biometrics, mood monitoring, life-logging, DNA sequencing, etc. etc. - please come and considering sharing your methods and results. We've got some new things to show also.

For details go to the Quantified Self Meetup page.

Alien Data

Gary Wolf

A column by Olivia Judson in today's New York Times touches on both scientific and literary testimony about the self-blindness of human beings. In "Wanted: Intelligent Aliens, for a Research Project," Judson points out that we are terrible self-analyzers, at least using the tools of our ordinary understanding and perception.

If there is anything living on Mars, it's going to be weird bacteria or the like, not little green men. Which is a pity. Because what we humans really need is a group of friendly, intelligent aliens to study us, and give us a report on what they find.

The problem is, in many respects it's difficult for us to study ourselves.

First, there are practical problems. It's easier, for example, to study organisms with much shorter lives than our own: when organisms have short lives, we can accumulate lots of knowledge about them in a single human lifetime. Hence, we know far more about bacteria, fruit flies and mice than we do about elephants, giant tortoises or sequoia trees.

Another difficulty: it's hard to do certain sorts of experiments. Many of the experiments we can do on fruit flies would be impractical or unethical to do on people.

But there's a deeper problem as well: it's hard for us to see ourselves in an objective way.

Judson gives some good scientific references and some literary testimony about human bias. She even quotes the 18th century novel Moll Flanders, whose heroine warns young women about the dangers of vanity. I appreciated this mention of Moll Flanders, one of the great instructive books about self-delusion, because from its author comes an answer to Judson's appeal: we ourselves can be the aliens we seek. Moll Flanders, of course, was not written by an elderly adventuress, but by Daniel Defoe, a novelist, journalist, and connoisseur of human blindness. Moll Flanders is a fictional human, and the words on the page - highly abstract instruments, when you think about it - let us see ourselves as if from the outside.
 
Among the papers Judson mentions is one by Emily Pronin: "How we see ourselves and how we see others." Science 320: 1177-1180. (PDF). This is a review article whose footnotes contain an entire education on human nature, and whose moral is that we owe each other a good deal of charity. We tend to judge others based on their outward behavior, not awarding much extra credit for any inward feeling of goodness, which - since we are not inside their head - we cannot perceive. Pronin asks us to perform novelist's trick of imagination; to indirectly perceive our own self-blindness in the evident self-blindness of others.

Pronin is a psychology professor at Princeton. In another paper, called "Doing Unto Future Selves As You Would Do Unto Others: Psychological Distance and Decision Making" (PDF), Pronin uses some experiments to show that when we make decisions about things that will affect us in the future, we discount the subjective feelings of this future version of our self in much the same way as we discount the subjective feelings of other people in the present. We choose good experiences today even if they will lead to bad experiences tomorrow, because we can't feel tomorrow. Our future self is a stranger to us.

Many self-quantifiers collect data because they believe it will have some value in guiding their experiments, and in informing their choices. Through collecting data, we fight bias. But by stepping outside ourselves, we also step away from what Pronin, quoting William James, calls the "warmth and intimacy" of our thoughts and feelings. This may be why so much personal data sits around unused. Such data may have implications, but it doesn't have rewards. I wrote about this problem in my profile of Piotr Wozniak, the inventor of a tool to make people learn things faster. Wozniak thought his software would take over the world. Instead, it is used by a relatively small number of highly motivated, scientifically oriented students; for everybody else, such a tool is not rewarding enough. Most people prefer something like Rosetta Stone, which doesn't really teach you very much, but which provides a great subjective feeling of learning.

How can we take the experiences of our future selves, and make them subjectively real for ourselves in the present? This is a great problem for self-quantifiers to try to solve. It is key to making our data meaningful. Novelists taught us to feel the inner experiences of strangers. Now, who will make the future feel real?

Daytum for Pretty Tracking

Gary Wolf

As a break from working through some of the heavy papers on mood metrics that I asked for and received, I started playing with Daytum, the place to make pretty, web-based charts from your data. Daytum is in an invite-only beta right now, but they are issuing invitations to all who ask as their capacity ramps up. The people behind the site are Ryan Case and Nicholas Feltron. Though the tools are simple right now the appeal is obvious.

Out of curiosity, I decided to take a look at the salutations used by people who email me. Most of my email arrives without a salutation. However, when there is a salutation, it is often either Dear Gary or Hi Gary. (If the email addressed to Mr. Wolf it is probably from a German magazine or a Nigerian businessman.) I've been curious about whether the Dear salutation is fading from email, as has been reported. My data say yes. In two of the samples, Hi Gary outnumbered Dear Gary by (very roughly) four to one. In the third sample, the ratio was much smaller; but this sample contained a lot of spam. In fact, a high ratio of Dear Gary in an email sample is probably a reliable signal that that the sample has not yet been filtered for spam. This point would be much clearer in the charts below if I did not include all the email that contained no salutation.

The charts below are static images, but they link through to the Daytum site, where they are much easier to see.