It’s all in the data…. until it isn’t

There are many things I can be legitimately be accused of. However overreaching in terms of anything even tangentially linked to exercise is not one of them.

Not unless it’s reaching over a honed athletes Broccoli-with-added-misery special lunch to get elbows deep in a massive portion of chips.

A consistent arc of my athletic ability is not the only issue with the digital narrative being pedalled here. I’m as skeptical of the hooky algorithms as I am of the source data. It’s kind of my job to be professionally cynical when some chancer thrusts ‘their latest visualisation’ into my scowling visage.

Data Lineage?’ I bark. They look confused. ‘Algorithmic veracity?’ I enquire with a look of utmost weariness. ‘Derivation rigour?’ I shout at their retreating form.  Standard data management lore holds ‘Crap in, Crap out’. That there shiny thing you’re showing me is mostly a pig wearing lipstick.

Honestly it’s all gone tragically downhill since the abandonment of the trusty slide rule for those new fangled pocket calculators.  This is of course not true. Well mostly not true anyway. You can show almost anything with data. Torture it long enough and it’ll tell you anything*

Combine this explosion of data with the Internet of Things (or Internet of Shit as it’s been memorably described) and you can’t move for data points sticking up grubby electronic digits for analysis. In Garmin’s case they’ve made the classic mistake of starting without the end in sight.

Garmin devices started with a tiny data set logging speed, distance and time. The development of GPS chips flipped that into basic navigation. But it was still a narrow set of attributes focussed on one individual.

Then Strava brought its segments, Zwift its gaming and suddenly we’re cheerfully sacrificing our privacy to lob random data into questionable public facing data stores. Tin hats not required, but the product of those services is our data being sold to third parties. We need to keep that in mind.

You may not care. You probably should. Google buying FitBit fills me with horror. They’ve already been caught red handed surreptitiously shipping individual medical data to health care companies. We all know FaceBook is basically an evil business model, but the likes of Google and the other massive content aggregators/search engines aren’t far behind.

Years ago the UK government attempted to quietly create a ‘UK patient database’ with those records being available to all sorts of nefarious third parties. It was called CARE.ID, and the reason you’ve probably never heard of it was due to it being stillborn after significant lobbying from anyone sensible who criticised it for the privacy/ethical shitshow it was clearly going to be.

We’ve pretty much arrived there by stealth. And while it’s hard to argue that AI/Machine learning unleashed on massive data sets could/might/hmm schmaybe create stunning new insights and breakthroughs for all sorts of conditions, it’s equally stupid to hand our data over to organisations whose entire value proposition is monetising that data.

Right sorry about that. Got a bit distracted. So this data arriving from cadence and power sensors is tossed together with heart rates, Vo2 guesstimates** and all that malarky to make an unwelcome metric salad.

Which is then liberally garnished in digital snake oil.  Ooh the shiny. To be fair, these kind of algorithms are predicated on some kind of structured training. It’s looking for a variance off a baseline.

My baseline was two weeks of not riding because I was too busy waffling myself stupid in Brussels, followed by all the self medication required to survive a wet week in Preston.

So no exercise to a fantastic three days riding in Hebden Bridge. Trouser tightness suggested this hadn’t assuaged the gluttony of the previous weeks. A glance at the scales nudging towards 13 stone confirmed that waistband metric. Right then things must be done.

Those things didn’t include riding outside. As outside appeared to be a crap CGI version of Waterworld. I’m okay with wet trails, less keen on riding through rivers. We’d done enough of the the previous week.

Turbo it is then.  Passed the 1000km of going nowhere slowly last Thursday. Not something I’m that proud of.  Most of those came from a mammoth 830k in a single month at the start of the year. No way I could face that regime again. So I’m mixing it up with a bit of virtual riding with my Bro, a few bastard hard interval sessions, a few group rides and no racing***

After dismounting earlier through the simple Newtonian principle of falling gently to the floor, that image was my reward. Unproductive. I know it’s bloody unproductive. It’s getting sweaty in a cold shed going absolutely nowhere. That’s pretty much the sodding definition of unproductive.

Well fuck it. I’m going to ride a proper bike tomorrow. Maybe I’ll ditch the trackers and just stop when I’m knackered. It’s not training is it? It’s dicking about in the woods with your mates. Good luck finding a metric for that.

Anyway I won’t be getting any more data devices for Christmas. I’ll be having that tin hat instead.

*I’m re-using that line from my latest Cranked article. What do you mean you’ve not got a copy? There’s loads of good stuff in it. And one of my articles.

**Calculating Vo2 Max from Heart Rate data is like me extrapolating the Severn tidal bore from standing outside and counting raindrops for a few minutes.

*** As a seamless transference of my real world racing prowess is going to beat the crap out of any remaining self-esteem.

 

Leave a Reply

Your email address will not be published.