We put up with too much from the tech industry. We accept two thousand dollar phones that break the first time we drop them, as though it’s reasonable to design sleek, slippery phones that are not robust enough to be dropped without harm. We surrender ourselves to the escalating arms (or eyeballs) race that sees every app trying to claim and monetise our attention without any consideration of how that affects us. We allow technology to force us to jump through hoops and shell out astonishing amounts of cash for access to our own data, or pay subscription fees to devices that we have already bought.
We suffer from a combination of shiny toy fever and an entirely hallucinatory dream that the tech industry is operating for our benefit, rather than to extract ever increasing rivers of money (and power) from us regardless of ethics, environment, or compassion. In the face of countless examples of sociopathic behaviour, our attitude towards tech remains bizarrely positive.
We really have been letting the tech industry get away with literal genocide for far too long.
Against the background of genocide, the issue of who owns our data and what right we have to access it seems minor, yet in a way it encapsulates the whole problem. Our phones, smart watches, and fitness trackers collect an astonishing amount of data about us, but they limit our access to that data in increasingly shameless and disturbing ways.
I recently wanted to put together a graph of my step count to show my surgeon, before getting both hips replaced. I have an iphone, and have been using iphones for a really long time, so it turns out I have a record of my daily step count since 2016 (which is probably the first time I owned a phone smart enough to record my steps). Using the health app on my phone, I could scroll through the average daily step count month by month, but could I download that data, or open it in a spreadsheet? Not easily!
Sure, Apple lets you “download your health data”. By which I mean you can “export all data”, not individual metrics. What it gives you is a massive xml file. Xml files are a bizarre pretence at structured data that is actually revoltingly difficult to wrangle, even for programmers. All but impossible to read.
There are various websites and apps of dubious provenance that I could choose to gift with all of my health data, in order to get back a file I could actually open in a spreadsheet, but if I don’t want to risk those, or write my own code to wrestle it into a useful format, then an xml file is effectively useless. It feels like an attempt by Apple to satisfy the letter of some law saying users should have access to their own data, without obeying the spirit.
Sure, they will give you your data, but good luck extracting anything useful from it!
I asked around to see how Android handles this problem, and discovered that extracting your data from an Android device can be even more difficult. There is no “export your data” button at all, but a google search informs you that you can use Google’s Takeout service to get a copy of the data. Kathy Reid generously navigated the process for me, and her comment was “hello! sorry for the delay – what a pain in the arse this was! Eventually got the data, but it wasn’t easy, and now I know why you asked for it”.
When a seasoned tech professional finds a process a pain in the arse, you can bet that most users will not bother.
Google Takeout gives you a zip file of folders with useful names.
Dig deeper into those folders, though, and things get… unnecessarily difficult. The Activities folder is full of tcx files – a form of xml used for fitness data. Unopenable by any software useful to the average consumer, but, to be fair, possible intelligible by other fitness devices. I didn’t try.
Let’s look at All data. It contains a slew of json files, with terribly helpful names like derived_com.google.activity.segment_com.google(25).json
Json files, while more usefully structured than xml files from a coding point of view, are not terribly wrangleable by consumer software, so good luck extracting your data from there. Or even figuring out which files contains the stuff you’re interested in.
Ok, let’s try the All sessions folder. Also json files, 2016-05-18T07_49_02+10_00_WALKING.json, but at least they’re helpfully named this time, with date, time, and activity type. A bit better, but still not terribly useful.
Our last hope is the Daily activity metrics folder. Hallelujah. CSV files – openable by spreadsheet software! Except… oh no. Each file contains a single day, and every metric recorded (or not) by the device. Which leads to a spreadsheet containing a lot of screens like this:
This might seem reasonable, until I ask you – how would I extract my daily step count from any of these files? Without the coding ability to wrangle the json files, I would not. Which makes the barriers to sharing this data with, say, your doctor or physio insurmountable for most people.
Scrolling down to the bottom of that folder, I did eventually find a file titled “Daily activity metrics.csv” which has the data aggregated, so there is a place I can find my daily step count. I suspect most people would have given up by this time though.
There’s a really important concept in the data world that doesn’t get enough coverage. It’s the principle of FAIR data – Findable, Accessible, Interoperable, and Reusable. It might sound a little abstract at first, but this issue with health data is nicely illustrative.
In the context of your own health data, findable means there should be a “download your data” option easily located. And it should be on each individual metric – so if I’m looking at my heart rate on the app, give me a download option right there. And another on my step count etc.
Accessible means once I’ve downloaded it, I should be able to figure out what it means and how to wrangle it. From a consumer point of view, I’d say this means I should be able to click on it and open it in software I can reasonably be expected to have, like a spreadsheet program (Excel, Numbers, or Google Sheets, for example).
Interoperable, again from a consumer perspective, means I should be able to pass the data from ecosystem to ecosystem, smart device to smart device, phone to phone. And I should be able to do it EASILY. If I buy a new device from a different manufacturer, I should still have access to all the data from the old device. I should be able to upload the data EASILY so that I can maintain an ongoing record of my own health.
Reusable in this case really means that we, the consumers, should be able to use this data for our own purposes. Not just browsing the slick displays on our devices, but downloading the data, graphing it, sharing it, and calculating our own statistics.
Certainly, most phone and smart watch users won’t want to wrangle their own data. But so many of us now wear and carry devices that automagically measure a vast array of things about our bodies. Imagine if doctors were trained to use that data for diagnostic purposes (and the training of doctors, and the modernisation of medicine, I am aware, is a whole other can of tangled wriggly things). Imagine if showing your surgeon the trend in your step count over time became normal, rather than startlingly bizarre.
If the tech industry continues to make people’s data unfairly unusable, we are condemned to ongoing ignorance, and deprived of tools we could use to monitor and improve our own health. If we actually made it easier to handle data, it would become less scary, and people might be more inclined to play with it.
Change, I suspect, requires legislation, which, in turn, requires a government and political class who understand this stuff, which is one of the many reasons I run a Data Science education charity. It’s crucial to our futures that we all become data literate. If we learn to understand data and are empowered to make change, we can all help shape the future, instead of being dragged, unwilling, into a future designed by two dudes having a cage fight.
You can donate to help power this work at givenow.com.au/adsei/