Interview with Dr. Rob Miller, Developer of rTracker App
A while ago I blogged about my search for the best self-tracking tools (link to part 1; , link to part 2), and one of the apps that came up in the final was rTracker. I have been using it since then, and have become a huge fan. rTracker now is a central tool in my Measured Me experiment – an attempt to capture and express my everyday life numerically. I usually try to stay as neutral and unbiased as possible in my reviews, but in case of rTracker, my engagement has evolved from being an enthusiastic user to becoming a “brand advocate.” I thought it would be great to finally “meet” a genius behind such an amazing app, so a couple of weeks ago I contacted Dr. Robert Miller, founder of Realidata Ltd. and developer of rTracker. He kindly agreed to answer a couple of questions for Measured Me blog.
Dr. Miller, thanks a lot for taking the time to answer my questions.
Thank you, Konstantin — please call me Rob. I’m really pleased that you like rTracker so much!
Tell us about your background, and how and when did you become interested in self-tracking?
My field is bioinformatics, which is the computer side of biochemistry. My primary interest is in protein structure prediction, but I’ve spent a few years on genetic sequence analysis, and my current job is taking me into chemoinformatics as well – the computer side of small molecules (e.g., drug candidates). My wife works in overseas development for the UK government, and we spent about ten years in Africa – where you may imagine I didn’t get to do a lot of bioinformatics! While we were in Zambia I worked for an internet service provider, where I started developing software tools to turn raw data from equipment and databases into useful information about what was happening on the network. I see now that I’ve been doing this same basic work ever since – finding ways to turn available data into information so that the trends become apparent (as in, ‘see the forest in spite of all the trees’).
It was also during the time in Zambia – in 2007 – that I started thinking about self-tracking. I wanted to see what kind of correllations there were between things – exercise, cigarettes, sleeping, alcohol, caffeine, weight, sex, happiness – were they influencing each other? Were there patterns that would show me how to get more of the things I wanted and reduce the things I didn’t?
What is the story behind rTracker? What inspired you to develop this app?
I could see that the core problem was in capturing the data, and it had to be fast, easy and super-convenient or it just wouldn’t happen. I started drawing up some designs for a flexible system for capturing day-to-day ‘life’ things, based on the premise that I could not prescribe ahead of time the input format. The one real requirement was that it be mine-able at the end — so free text was ‘out’ and structure was ‘in’, but at the same time simple numeric values just wouldn’t be enough (“I’m feeling about 7.6 today. How about you?”). The other piece that came up around this time, after a friend said my idea made her think of her personal information getting out on to the Internet, was that I had to address privacy at some level — even if it was only so that I could demo the app to someone without opening all of my personal life before them.
After we moved from Zambia I decided to start really working on the idea. It didn’t take too long before I recognized that all the platforms I was considering – iOS, Android and Symbian – could not be supported with a single code base and variations for the user interface pieces. I eventually decided to target the iPhone first, so bought a Mac Mini and an iTouch and got to work. At the same time, I started doing more tracking with paper so I would have real world data to feed in.
Do you keep in touch with users? What kind of things do people track with your app?
Yes, though I mostly only hear from users for feature requests or the occasional bug report if they can’t find a work-around. I’d love to know more, like what trackers they’ve put together or whether use the functions or the privacy part. A friend of mine – customer number one – tracked his dog’s dimensions as it was growing and now his exercise, but other than that I think I’m be the only self-tracker I know! One feature request that I implemented was to add ‘per-calendar-(day,week,month)’ as a function endpoint, so that the user could watch e.g. how much coffee they’d had that day instead of ‘over the past 24 hours’ as the example Drinks tracker calculates. I also get hints from some of the reviews – like using the multiple choice items for mood terms, and measuring medical parameters during a hospital stay.
I like the tracker ideas you show in the screenshots on Measured Me. I find mood measurement to be particularly difficult to do objectively, so there’s a lot of space for coming up with the right questions to tease out honest measurements.
Do you currently track anything? What kind of apps and gadgets do you personally use?
Yes, I’m always tracking stuff. Most of my own trackers are the ones that have made it into the included example set, or are available for download from the website (note from Measured Me blog: rTracker app has an amazing feature – you can add other people’s questions, called trackers; check out and download sample trackers here; I will be making my own trackers available for download soon, too). I usually create them with just the data collection parts, then work out functions later to look for useful information. The Sleep tracker was a good example of that — for the longest time I could see the perfectly predictable pattern that I got less sleep during the week and more on weekends, but when I started averaging sleep duration over the preceding month it became a flat line just under 7.5 hours. The nice outcome from that was realizing that ‘eight hours’ wasn’t actually what I needed – and now I have an extra half hour in every day without worrying about it.
Recently I’ve been tracking my household electricity and gas usage – just read the meters every Saturday, and rTracker calulates average consumption per day since the last read; much more useful that looking at the bills, where the company just estimates and then adjusts with a true reading only once or twice a year (note from Measured Me blog: Dr Miller refers to another awesome feature of rTracker – function: you can perform real time simple calculations and data transformations right inside the app). I also started trying to track my commute – I’m fascinated by how traffic changes as a function of time (if I start 5 minutes later it seems the commute takes 10 minutes longer) or day of the week – but haven’t worked out how to process and display it usefully yet.
In terms of gadgets, about my only one is my ‘Weight Watchers’ scale which claims to report my percent body fat and other characteristics. Don’t know if it is accurate, but the measurements do vary over time in a logical way and provide a bit more insight than just my weight and waist circumference.
Last year, I wrote a post about how the “perfect” app for self-tracking should look like (link to part 1; link to part 2). I would love to hear your thoughts on that. You mentioned in your email that you plan to continue improving rTracker. Do you have anything specific in mind?
Obviously we are on the same page with respect to versatility around setting up trackers and the items within them, so not much to add there! I (and many other developers, I expect) would love to be able to seamlessly capture data from other sources, like other apps, but we are pretty much blocked from this by the ‘sandboxing’ security on today’s mobile device operating systems. Most hardware add-ons will come with their own (sandboxed) software – and even if they do open up their interface specifications, its a gamble whether supporting a piece of hardware will generate enough sales to pay back the cost of buying one to work with. Supporting external web-based interfaces is more feasible, but their interfaces become another thing to keep up with — and often their income is from serving ads on those pages, so they won’t want to lose eyeballs to my app.
At least rTracker can share data through CSV files as a currency for both data export and import. As I noted above, I did lots of pencil-and-paper tracking before rTracker was usable – so to get that data in, I could type it into a spreadsheet and load it through iTunes.
I fully agree with you about being able to track and review your data without being online, and that feeds into my feelings around security of tracked data as well. Menstrual cycles, drug use, spending habits, sex life, depression, I don’t know what private things you might want to track — but if I were you I wouldn’t want it sitting on a server waiting to be stolen or sold. rTracker doesn’t do real encryption (another can of worms for developers), but if you’re sufficiently worried about that you will have sorted out security for all the data on your phone.
In terms of rTracker’s roadmap, the next release will expose functionality around sharing – emailing CSV files as you suggest (and I promised one of my users a few months ago), and more. After that, push notifications are high on my list (e.g. being queried at random times may help the subjectivity problems of measuring mood), and I’m always coming up with graphing features and function capabilities to add. rTracker’s not making me rich (yet!), so the developments are all around what I want as a user – a ‘labor of love’ as they say – but this also means that I don’t get to work on it as much as I’d like.
Do you follow developments in the Quantified Self movement? Are you a member of any QS meetup?
I came across the term ‘Life Logging’ after I had been actively working on rTracker for several months, and only more recently did I hear of ‘Quantified Self’. I live near London now and will try to catch up with one of those meetups one day; maybe a five minute talk on rTracker would be a good way to introduce myself?
There is a lot of emphasis in QS on passive measurement and objective data. Do you think self-reports and subjective data still have place in the tool bag of self-experimenters and self-trackers?
Clearly passive measurement would be the ideal, and objectivity would be brilliant – though for mood it might need a brain scanner! I love the idea of developing rTracker to interface with hardware that captures all sorts of life data, but I just don’t think it is realistic yet beyond hardware already on a device like the iPhone camera, GPS or accelerometer — and continually watching these will likely impact battery life pretty badly. This means more tools for each passive/objective data source, and the result is straight back to the gadget / app proliferation, the ‘I know what you want to track’ thinking, and the inability to bring different variables together that I wanted to address with rTracker. So yes, at least until we get a step change to open protocols for capturing and sharing QS data, I’d say self-reporting is the only game in town for now.