I’ve been “life logging” off and on for the last ten years and I’ve gotten a lot more interested in and serious about it in the last two years since devices like the Zeo sleeptracker and Moves for tracking movement have become available. Back in July I built Lifetracker, an app for tracking any arbitrary variable in your life. Building and using this app has been enlightening. Now I’ve spent the last month researching the Quantified Self space and I’ve set out to to truely understand the space and get familiar with all the tools and companies out there. Every. Last. One. My iPad is literally out of space because of all the apps I’ve been trying out.
The space, as it is now, is made up of inputs inputs for collecting data and outputs for viewing and interacting with that data. Then there are services, apps, devices, and APIs that act as both inputs and outputs. Most things attempt to serve as both inputs and outputs. For example, Fitbit and Reporter both track and show data. Some devices also analyze data and find correlations.
This is all great, except that the space is fragmented. There are dozens of devices, apps, and services out there to choose from. This fragmentation causes problems: you cannot compare data between different inputs and you have vendor lock - you cannot easily switch between different inputs and expect to keep your data in a single place and format.
And I don’t expect this fragmentation of inputs to go away any time soon. Nor do I think it should; It means that there are a lot of different people with all sorts of approaches building all kinds of devices to help you quantify your life. That’s a good thing; it’s competition.
Outputs are similarly fragmented: there are dozens of services that exist for the purpose of aggregating, organizing, comparing, and visualizing your data. The Apple Health app and Exist.io (among others) are trying to solve this by providing a single place to view your data - but even these have problems. They still ensure vendor lock-in and they are both private companies that are now in possession of massive amounts of very sensitive, personal data. Additionally, if you wanted to add data from a device that is not supported by Exist or Apple Health, you are out of luck.
So here is my proposal.
1. We need a standardized API and data format.
It might be a bit early for this, but it’s time for the industry to start thinking about it. Human API and Fluxtream seem to be the two leaders in this; they both have very well thought out API’s for accessing health data and I think they are a good start. But I think we can do better. We could have something as standardized as the HTTP protocol allowing anybody to develop new inputs and outputs that are totally interchangable. We no longer have to rely on input device manufacters like Fitbit to also be good at creating pretty charts and doing data science. Device manufacturing and data visualization are very different. Unless you are Apple, I don’t expect you to be good at both the hardwire and the software.
2. We need a modular, open platform for tying everything together.
I love all of these services that are aggregating data, but I’m not satisfied yet. I want a platform that I can attach my Raspberry Pi or Moves data to and then write my own correlation finding algorithm for. I want something that I can write a plugin for. I want a hackable, open source aggregation tool with a vibrant community and lots of plugins.
The Quantified Self space is a geeky space full of hackers. I want to build a platform that is intuitive enough for any non-tech-savvy person to be able to understand and use, yet open and available for any programmer to open up and write plugins for. If you want to design a new pie-chart interface, why not? If you want to design a new a correlation-finding algorithm, why not? If you want to hook up your Raspberry Pi and write your own import script, why not?
So I propose to build exactly what I have just described: an open-source, hackable tool for aggregating and displaying Quantified Self data. I want it to be modular enough that anybody can write a plugin for it. Something that anybody can build a new visualization page for if they want to. A tool that anybody can download, build, and run on their own laptop or computer or server. A tool that lets someone truely own their data. A tool that is entirely free and that doesn’t require you to trust and depend depend on a private company that could go out of business or sell your data for marketing purposes.