andyhickl

Andy Hickl is an innovator in computational linguistics, natural language processing, text analytics, machine learning, and large-scale data processing.

A quantified life, not a heartless one

Tiffany Jenkins, writing in the Scotsman, has a problem with the quantified self: it has no heart.

Here’s an excerpt from her evocative piece:

We are in danger of outsourcing the essential examination of life to a computer programme and reducing our wonderfully messy reality to numbers and a spread sheet. Relying on data to interpret our lives means we are giving up on doing it ourselves, through conversation, debate and private thought, through dreaming about what could be.

Needless to say, I found this provocative enough to blog about.

Jenkins makes two arguments in her article. One I completely agree with, the other not so much. Let’s start with our differences, shall we?

First off, Jenkins argues that quantification necessarily gets in the way of living life. In her view, paying attention to personal data can only inhibit the enjoyment of our everyday experiences.

We should delete the data and think for ourselves. Life is for living, not logging. Nothing else adds up.

For Jenkins, collecting data causes us to become too results-oriented, too interested in outcomes, as opposed to experiences. We end up focusing on the wrong things: how many steps we took today, as opposed to what we saw along the way. She even hypothesizes that too much personal data can cause us to become disconnected from our humanity, like some sort of modern-day Thomas Gradgrind.

I don’t think it has to be this way. Technology has advanced to the point where lots of personal data can be collected with little intervention on the part of the user. My Jawbone wristband is a little like Santa Claus: it knows when I’m sleeping, it knows when I’m awake, it knows how active I’ve been during the day. My lifelogging app, Saga, knows where I’ve gone and what I’ve been doing. Soon there’ll be cameras (like Memoto and Authographer) to capture everything I’ve seen and other devices to capture everything I’ve heard.

But the key here is that none of these impede me in any way. I’m not missing out on anything because I’m lifelogging. I’m not doing anything differently.

(In fact, that’s the point of lifelogging apps like Saga: to capture your real life. Lifelogging doesn’t motivate me to go to fancier restaurants or stop more often at scenic vistas or snorkel in Caribbean lagoons. But if I do, I’ve got a resource that can remind me months from now that I did (or didn’t do) all of those things.)

Jenkins’s second argument is that quantification is no substitute for good, honest introspection. She writes:

The unexamined life is not worth living for a human being, observed Socrates. But is this the sort of self-examination he meant? I don’t think so.

Firstly, much of life is not meaningfully understood through data. Take happiness. An app may count the number of times I tick that I feel happy in a week and note that it’s infrequent on a Monday morning but that it increases on a Friday night. But it cannot explain why I might not want to feel happy, or that feeling unhappy is a spur to get things done.

Only I can make sense of my emotions, with personal reflection that is not reducible to numbers. The inner life cannot be understood through quantification.

While I’d argue (obviously) that quantification should play some role in understanding our inner lives, I’m largely in agreement. The crop of personal tracking devices and apps available today largely don’t understand how the metrics, events, and indicators they record impact us as humans.

My step counting app doesn’t know that I’m taking fewer steps today because my shoes are hurting my feet. Conversely, my lifelog can’t understand why a frenetic trip through the drive-thru for lunch isn’t as fulfilling as a leisurely meal at the Four Seasons on a sunny day.

But that’s not to say that’s always going to be the case.

Lifelogging apps, in particular, are opening the door to more empathetic interactions between devices and humans. A lifelogging app like Saga can already know when I’m in a strange city, or when I’m stuck in a endless stream of meetings, or when I’m slower than usual getting out the door and getting to work. If we also add in information about what music I’m listening to, how well I slept last night, or details about my last workout, it becomes clear that my phone could know a lot about my mental state, mood, or current interests — and adapt my user experience accordingly.

Of course, that’s not the same as enlightenment. But it’s a good start.

Wearables, apps, and healthcare

Today, it’s easier than ever to capture information about your health. Just turn to a wearable. Or a smartphone app. Or both.

There’s lots of great, sexy-looking wearable hardware. My Shine keeps track of my activity level. My BodyMedia Fit lets me know how many calories I’ve burned and how well I’ve slept. I’ve also got a Jawbone Up, a Nike Fuelband, a Fitbit Flex, a Basis band, a Withings Pulse — and an insane number of chargers to keep track of.

Despite all the attention paid to wearables, it’s great to see that smartphone apps are holding their own. Moves and Argus are great at using my iPhone 5S’s motion chip to keep track of how many steps I’ve taken. Azumio makes apps that let me use my phone’s camera to track my heart rate and stress level. And that’s not to exclude automatic lifelogging apps, like Saga and Chronos, which are capturing data about my behavior, including where I’ve been, what choices I’ve made during the day, my mood, and so much more.

This is just the beginning. Analysts are predicting that we’ll see even more growth in the months ahead. ABI predicts that more than 100M personal tracking devices will be sold each year by 2016. There’s reasonable expectations that personal tracking will be a $12B industry by 2020. Investments have been huge, as well: Fitbit raised more than $40M this year, BodyMedia was bought by Jawbone for more than $100M, and personal informatics projects on Kickstarter have booked more than $6M in revenue alone in 2013.

While it’s clear these apps and devices have attracted a loyal following, most of us are still trying to figure out what to do with all of this data we’ve collected. Unfortunately, there’s doesn’t appear to be much agreement on what the killer app is for personal informatics.

This weekend, I organized a panel at MedicineX looking at one possible application domain for data captured from wearables and apps: healthcare. I was joined by Sonny Vu, CEO of Misfit Wearables, Ivo Stivoric, VP Technology at Jawbone, Jon Kiehnau from Spree, and Ted Tanner, CTO at Pokitdok. (You can check out the HD stream here.)

Our going-in position was pretty simple. Wearables/apps provide unparalleled access to continuous streams of health data that’s available nowhere else in the market. They’re cheap, reasonably accurate, and don’t require the user to have to remember to self-report.

These data streams are particularly exciting in that they enable the same kinds of analytic effects usually reserved for enormous “big data” sets. Since they’re (presumably) always on, wearables and smartphone apps can capture sparse events that might not be recorded by sampling. The completeness of the data doesn’t just make it easier to visualize or to spot trends — it also provides a richer feature space that can be used to build models or support inference. In most cases, it doesn’t matter that the data comes from only one person; these “little data” sets can hold just as much value as the traditional “big data” sets that we’ve all come to know and love.

There are plenty of hurdles to overcome to actually using this data in healthcare practice, however. Quality is a biggie: the data’s gotta be good enough to be used in a clinical setting. We also have to be careful that we don’t drown practitioners in a tsunami of personal data, as well. Data streams need to be handed off to the practitioner in a format that’s easily consumable, searchable, and analyzable. Finally, these continuous data streams need to be contextualized as well. Health care practitioners need to know what external factors could have caused a rapid spike in blood pressure — or true diagnosis can’t happen.

While we do have plenty of work ahead of us, it’s clear that wearables and apps are bringing about a “little data” revolution that has the chance to really impact the healthcare consumer - practitioner relationship. If providers can leverage personal informatics provided by the patient in their diagnosis, there’s no doubt in my mind that we can build stronger relationships between consumer and provider based on real data.

This is my calendar.

The yellow/orange blocks represent meetings that were scheduled by humans. The blue blocks? My exact whereabouts as recorded by robots Saga.

No, you can’t have it yet.

This is my calendar.

The yellow/orange blocks represent meetings that were scheduled by humans. The blue blocks? My exact whereabouts as recorded by robots Saga.

No, you can’t have it yet.

Saga + MyFitnessPal = A better me

I’m really excited about Saga’s new partnership with MyFitnessPal. Let me tell you why.

We think that understanding your past activity can have a real impact on your overall health, well-being, and fitness level. If you’re like me, you use a whole host of services to track your progress towards your fitness goals. I use a Jawbone UP to learn about my sleep patterns, a BodyMedia armband to know how many calories I’ve burnt during the day, MyFitnessPal to stay on top of what I eat, an iPhone app to track my steps — and Saga to understand where I’ve gone and what I’ve done.

Why? Well, for starters, Saga gives me an important perspective that no other service really does. Over the past year, I’ve learned that my weight goes up when I work more than 50 hours / week — or when I average more than 4 meetings a day for a given week. Saga helps me understand when I’m working too much — so I can remind myself to get a workout in (or at least watch my nervous eating).

Saga’s told me that I’m more likely to go for a run when I’m traveling — so I should definitely find space for my running shoes in my luggage (and scout out a good route before I leave).

I’m more in touch with my fitness — and what make me fitter — because I use Saga. And now that my MyFitnessPal data can show up in my Saga lifelog, I’m fully expecting to be an even better me.

No, Saga may not be for everyone. But we’ve got an unbeatable privacy policy which I can summarize as follows: we will never ever ever ever sell your personal data. Really. You can see it here.

Let me elaborate. We believe the data we collect from your Android or iPhone is yours. Want us to delete it? Hit the button in our app, and we’ll do it. Want to download all the data we’ve recorded? Contact us, and we’ll give it to you. No questions asked.

We have no designs on being Big Brother. Or even Little Brother. Just like MyFitnessPal, we want to be another service that you can use to help you understand what makes you awesome.

Any questions? Hit me up at andy@aro.com. I’m happy to help.


A version of this post was posted to Mike’s MyFitnessPal blog on August 2, 2013.