The Quantified Self and Humanities Best Friend.

I never thought I’d be the kind of person to “do” a 5k. In fact, had you asked me a month ago, I’d have told you that Americans don’t understand the Metric system, and that 5k runners were clearly invading Canadians in disguise. And yet, two weekends ago I found myself standing at the starting line of the Santa Paws 5k surrounded by 300 invading Canadians and their four-legged best friends. Santa Paws is the annual fundraiser for the Wake SPCA, here in North Carolina. Of course, I couldn’t just walg (walk-jog) the 5k, I had to turn it into an excuse to play with new technology!

The Quantified Self.

For awhile now I’ve been fascinated by the idea of the Quantified Self. In short, the idea is to log one’s activities, locations and events. The hope is that more data can lead to better decisions. Some of this has been going on for decades. Keeping one’s day-planner, or diary updated is likely the analog beginnings of the Quantified Self. The innovation of our generation is found in passive collection. Put a Fitbit, Up (or soon, Apple Watch) on your wrist and your daily sleep and activity is automatically recorded. In the age smartphones with motion co-processors and clever apps we can even move beyond automatically gathering data, to proactively prompting actions. Everyday, ’round about 5pm RunKeeper lovingly(annoyingly?) tells me “Remember, you used to think this was a fine time to go walk? Want to go for one now?” Our access to technology has raised the bar from passive collection of data, to actively prompting and encouraging better decisions. (And yes, even I have to admit that more activity is a better decision for me.)

The Problem

Kevin and Lilo rest after the finish of the SantaPaws 5k

Kevin and Lilo rest after the finish of the SantaPaws 5k

Lilo is my adorable 45lb Carolina dog. She’s ferociously smart, loves Babyfriar and is the secret to cheap, boundless energy — if we could only figure out how to harness it. I’ve worried that she’s not as active as she needs to be because her brothers are decidedly lower energy critters. She needs more exercise. Right about the time I signed up to do Santa Paws with Lilo, I found out about Whistle, wifi enabled ‘Fitbit’ for dogs. Whistle uses a Bluetooth connection to your phone to identify who the Pup is with during activities and can differentiate between a walk, playtime and sleeping-in-the-most-uncomfortable-looking-position-on-the-couch. A quick trip to the pet store and Lilo was Whistle’d. Not only are we quantifying ourselves, but our best friends as well. Indeed, with the RunKeeper app providing information like distance walked, and the current and average pace while recording GPS locations; it’s possible to not only quantify any given walk, but also use that data to, for instance, slowly increase pace and distance for training purposes. The problem, however, is that the RunKeeper data is siloed off in the RunKeeper app while the Whistle data is hidden off inside the Whistle app.

Enter Salesforce

I believe Salesforce provides an ideal platform for self quantification. It’s api’s provide a rich environment for integrating and aggregating data from a myriad of sources. And so a project was born: A real-time updating map display of Lilo and my progress on the 5k course with GPS data from RunKeeper, step information from Fitbit and activity information from Whistle.

Building it out

As with many a side project, the design was simple on paper, but proved rather challenging to implement. I wanted an app that would:

  • Collect data from Fitbit
  • Collect data from RunKeeper
  • Collect data from Whistle
  • Display the data on a Map
  • Update the map in realtime as new measurements are recorded by RunKeeper
  • Expose all this in a nice way to my supporters so they could see the race progress.

The design

My starting, back of the napkin design had me using the Streaming api to deliver updates to a public, authentication-free site. Unfortunately, I quickly discovered I can’t use the streaming api on an unauthenticated site — sadness — so instead I wired it to periodically pull new records via javascript. I also discovered that, despite my best intents, hooking Runkeeper, Fitbit and Whistle all up was a too-tall task for the 6 days I gave myself between hatching the thought and actually walking the 5k. What I ended up with was an app that received, and interpreted GPS and walk data from RunKeeper and displayed it in near-realtime on a map. You can view a replay of our 5k walk here.

The data is populated from RunKeeper via a Heroku based Rails middleware app. As I have time, I’ll flesh that out with data from Fitbit annotating each marker with the number of steps taken since the last marker. Unfortunately, Whistle has declined (as of yet) to publicize an API, leading to a rather hacktastic, unofficial api that is currently broken (or my charles-proxy-fu is weak.)


Ever look at a cloud and think “Oh, Hey, it’s a [Star Destroyer | Tardis | Puppy eating a Tardis! ]?” If so, you’ve experienced apophenia; the experience of identifying patterns or meaning in seemingly random data. Our brains are really good at finding such patterns, but the information isn’t always presented in ways we can easily process. This is where analytics steps in, helping us visualize data and in so doing, understand what it all means.

This is the real reason the Salesforce1 is such a great aggregation platform. It’s suite of analytics tools — from reports, dashboards and the analytics api that can power D3 based visualizations like the one’s Christophe Coenraets has blogged about — are unparalleled. After we’ve moved to 30 hour days, and I have more time, I’ll update the maps page to show some D3 based analytics showing the number of steps and elevation gain. That’s a start on visualizing how my pace slows per % of grade uphill and demonstrates that running downhill after your dog is still faster than tripping and rolling downhill after your dog.

1 comment

Leave a Reply

Your email address will not be published. Required fields are marked *