01100111 01110000 01110011 01110010 01110101 01101101 01101111 01110010 01110011 00101110 01100011 01101111 01101101
You probably don’t know what the above says, other than a bunch of binary numbers. The numbers are below:
103 112 115 114 117 109 111 114 115 46 99 111 109
These are, in fact, ASCII codes. They say “gpsrumors.com”. Refinement of raw data is what makes it useful. I think we can all agree that a binary interface is almost useless to humans, although some older computers did display the binary using lights. What I’m seeing though is that very few people realise the same is true of fitness data. Raw data is useless unless you’re very experienced, and even then it’s time consuming and error prone to use in that form.
This diagram is the information pyramid. It shows how the processing of data makes it more useful. Raw data at the bottom is just that: data. A bunch of datum’s (the single form of data) gathered together. In running this would be a location, current pace, or cycling maybe current power in the right pedal. Not useful. Gather these and process them though and you can create information. That information might be heart rate zone during a workout phase at a given pace. Much more useful, hence the term “information” is used rather than “data”. This information is valuable to humans. We can go a step further, though, and create knowledge. This would be using lots of information to create something actionable. For instance how is my heart rate at a given pace over time. This gives me the knowledge of whether I’m improving or getting worse. If we have enough of this knowledge we may even be able to create wisdom. At these heady heights it may be the case that we can learn how to make something happen.
In data analytics, we often show the below diagram. This is kind of what we’re going for any time data is involved. As you move along to the right things get harder, but value goes up exponentially.
On your watch, you’ll see the “what happened”. As you run it tells you how fast you’re going, where you are, climb rate etc. Really fancy watches may also give you some more insight, such as telling you your steady state workout produced a certain outcome for fitness. The wider app and platform is (in theory) there to dig into things further and take you up a notch.
I say “in theory” because as we all know they don’t really, do they? Garmin have more fitness data than just about anybody on the planet. You’d think they could get a handle on what works and what doesn’t, and be able to properly feed back to us. They won’t. Not yet. Until they absolutely nail this technology they have a problem. Their business is driven in part by coaches and trainers. And these people need you to need them in order to survive. And you’ll likely buy whatever they use, because they use it. The far right of that graph up there is where machine learning and AI start to appear. We’re already there today from a technology standpoint (trust me, it’s not even that hard!). What we need to do is spend time and energy training the bots, and testing them.
What’s that, you say? No way to replace a human? If a human can make a decision in less than a few seconds, a machine can make the same decision faster. Coaches have a set of rules in their heads which they made through training and experience. Machine learning not only does this too, it does it faster and more accurately with less superstition and bias, AND it can use billions of data sets rather than the few hundred a human could handle.
Anyway, that’s the future. That’s when we won’t need the information display because we’ll be told what to do. Not been for a run in a month? Your watch will beep loudly and suggest you get off your fat ass. Probably Fenix 6 or 7 will do this. For right now, you need to interpret your data or get a coach to do that for you. And this is where we turn the data into information.
Graphs are easy, right? Just a few numbers, a couple of axis.
Here is the Garmin heart rate graph. I didn’t have a heart attack, don’t worry, I just own a Fenix 5. Imagine the gaps aren’t there. This is data. The only way this would be worse would be to have the raw numbers printed in a table. What am I supposed to glean from this image exactly?Garmin also allow you to switch to a different tab and see how long you spent in each zone. Again, I’m not sure how this really helps aside from knowing it was mostly a zone 3 run.
Again with the pace graph, I can barely read this and it certainly doesn’t allow me to see my pace easily for the different intervals. I can hover over with the mouse, but then why bother with a graph? Next we have the Suunto graph (different run). It’s spikier. That’s literall all the value I get from this compared to the Garmin. It also looks cooler, like someone with a design degree was involved at some stage. Garmin is all engineer all the time. No designers in Kansas.
Suunto also include a bar chart for zone time. You don’t have to switch tabs, just scroll down so there’s a little more value seeing them side by side. Lots of wasted space, but it does have a certain charm.
And finally Polar. It’s important to know that Polar are very sports science focused and they also happen to make some devices to enable that. I feel as though Polar may have some designers, but they weren’t allowed much input here. The engineer was told exactly what to produce. The real input, I assume, came from someone who does sports things for a living. This is not data folks, this is information. At a glance I can see which zone I was in, and how well I was keeping within that zone. I could see if I was at the top of the zone the whole time, or near the bottom, and I can see instantly how long in each next to the graph. Someone thought this through.
Sometimes it’s handy to see HR or some other metric against pace. Garmin allows you to expand the graph to do overlays.
Tables can be made more interesting too. Tabular data isn’t the most exciting, but sometimes it is useful. Here is the tabular data on Garmin of my recent swim. Here we can easily see the set totals but also expand to see the set data. We can also very easily see rest intervals, something quite important for swimmers. Swim repeats are one of the most used methods in swimming, and Garmin helpfully supports the “repeat on” function to allow you to start your next lap at a given time. For instance, repeat 100m on 1:45 would mean that if you swim your 100m in 1:30 you get 15 seconds to rest. As the set goes on your rests get shorter as your times get slower. Seeing this against HR information would be critical to determining efficiency of your technique in the pool (way more important than fitness or strength in swimming).
Garmin also give a nice graphical display of simplified stats. I like this view a lot for quickly seeing what I did on a whole swim before drilling down with the data.
Polar shows a very similar display to the Garmin table. They don’t have graphical mode, but they do have two “killer features” and one drastic disadvantage. Firstly my optical HR works in the pool, so I have that info. It seems accurate too. Secondly, I don’t have to push anything in the pool. Polar detects my stroke, my lengths, my laps, and my rests. My rest timer was identical to the garmin on the other wrist where I pushed the lap button. I was very impressed. It lacks a repeat timer though, so for swimmers it’s not quite there yet. Right now I can rest for 20 seconds but I can’t repeat on 1:45. There’s a very big difference, and you only need to watch swimmers staring at the wall clock to realise they use the repeat timer exclusively.
Polar are doing great things here. They consistently make information out of data to help me on my way. After I cover on-watch information display I’ll come back and cover the knowledge level aspects of each platform. Again, Polar are out front on the knowledge front (albeit with some bugs…). I’m finding the Vantage V system great for telling me what’s going on between workouts (albeit with some bugs…). On the watch face I can actually have a symbol to say whether to work out or not in a simple easy to use way. That’s a win, and that’s the future or wearable tech. Garmin tells me some stuff too, and one day I’ll investigate how to interpret that jibber jabber. Until that day it’s like I’m a cat reading Shakespear, and that’s a fail. That’s the past.
It’s only fair to point out that although Polar do well in this post I am having several major issues with the Vantage V right now so can’t recommend it as your only device. It recorded a 5km treadmill run as 2km today, and since I got it, it’s told me not to train or I’ll get injured. I did get injured, but that was due to running in the dark in an unlit lane and twisting my foot into a pot hole. Add that to the “oops” I got recently when it crashed and I have to say wait for the next firmware before buying, or make sure you have two devices.
I’ll be adding a second post soon to show you the data/information discussion on the watch itself. Suunto really shine there, and Polar are going places too. Garmin let the engineer have too much say again, and rely WAY too much on people filling their gaps with ConnectIQ. There are still gaps.
Thanks for reading, add your thoughts and comments below, I’m sure you’ll have plenty as usual and remind me of things I should have covered which I’ll add as requested.