I have now finished the baseline phase of the residency. I’ve got good sleeping data for that period, with near continuous sleep mode and night time movement. I’ve got a few nights of heart data, hopefully that will become more continuous now. I’ve got solid weight and fat percentage data, only missing those days I’ve been travelling.
I’ve got near continuous daytime movement data, with a few logs of heart data onto the trip to work. Again, I hope to get a more continuous record of heart data now that I’ve upgraded the heart monitor. I’ve got slightly patchy data when it comes to the size measurements and sleepiness questions, but that is being more reliably recorded now.
What I haven’t got is a public drop of this data, which is unfortunate, but I’m working towards that.
This week I upgraded my heart monitor to one that allowed me to get detailed data of my heart rate, more about this in a separate post. I started taking semi-regular measurements of my girth. I also started recording results of the Epsworth sleepiness scale up on my Google Drive. I got a good start on the design of my installation, and have started communications with the appropriate staff at The Edge about how to get it implemented. I also had some external confirmation of the data format I’ve decided to use, which is always a good thing.
Most importantly I finished my Basis phase this week, and begin by Mental phase next week. For this coming week, I have a few posts to write, I have to make a first drop of my data set, and I have to continue looking into the installation.
In terms of oddball things, I’ve been wearing my sleep monitor head band so regularly at night that during the day I feel like I’ve lost something. I’ve even taken to wrapping my shemagh around my head bandana-style to stop myself freaking out so much.
This week I continued with with my sleeping baselines, hopefully resolved the heart monitor issue, and started work on one of the deliverables of my residency: the interactive visualisation.
I have upgraded my heart monitor, hopefully to a model that actually can give the detailed logging I’m after; I’ll find out for sure tomorrow when I upload the data from the watch to the computer. I’d really like to get some baseline data of what my heart is doing during sleep, exercise and work.
Next week I aim to start measuring:
- Some subjective things: how tired I feel in the morning and during the day.
- What I’m eating.
I also need to look at another one of my deliverables, the publishing of health data. I’m happy with the format I’ve chosen, but I am yet to choose a backend to store the information. At the moment I’m leaning towards storing it as a simple file until it gets too large and I’m forced to store it in a proper way.
I would like to get the interactive visualisation of my collected data up as quickly as possible. So it has to be quick and easy to put together, yet eye catching. The installation also be able to grow as the project does. I really want the installation to be up for as long as possible to increase the number of people who know about the residency.
Essentially, my solution is a fancy projection surface, relying fundamentally on the form of the Vitruvian man. I have created a silhouette of the Vitruvian man, combining both positions into one. There are three planes used as a projection surface, the body, the ring and the background. Note that I’ve tried to incorporate the original bounding circle and square into the final shape.
A projector will be placed in the back of the installation. The raw nubmers and graphs will be projected onto the back surface, while the visualisations of this data will be projected onto the body surface. The outer ring will be the interface, divided up into segments for each type of measurement I’m collecting.
The interaction with the interface will be done via a Leap Motion device, which allows hand gestures to drive the interface, up/down motions will rotate the selection bar, while left/right motions will turn a particular segment on and off. Once the measurements to be displayed are selected, the UI bar will turn into a date selector. Once the time span is selected, the data will be projected over the two surfaces.
At the moment I’m seeing it mostly being built out of clear perspex, with white tape for the projection surface, and black card to block projection over the raised areas. There are lots of questions, mostly around how deep I have to make it for it to be self standing.
I’m writing this on a plane to Perth after a complimentary wine, so this might blather on a bit. This week I got to grips with a few storage issues and started getting baselines for some of the data I wanted to track. I managed to muck up purchasing of two of my gadgets and discovered that the owner behind a third appears to be going out of business.
This week I started using the scales, the pedometer, the sleep monitor, and the much dreaded tape measure. I also started using a heart monitor, but as that gadgets article will show, I purchased the wrong one. Next week I hope to start tracking my food intake, probably by taking pictures with the mobile. I’ll also start logging my general mood, with a questionnaire. I’ll also hopefully upgrade my heart monitor and start tracking that as well, at least for sleeping. As mentioned earlier I’m travelling during the weekend, which presents some problems for the continuity of the project.
The Fitbit and tape measure travel quite well, though the fitbit, which is a wireless transmitter and thus must be switched off, has a rather odd way of being switched off: it must be connected to power in charging mode in order to be switched off. The sleep tracker is a tiny bit more cumbersome to carry, as it’s charging dock has to be taken as well. The scales, unfortunately, just aren’t very good as carry on.
The heart monitors I’m looking at are the same used by gym junkies and elite sports folks. There’s a strap around the chest and often a wrist band. The basic idea is that there’s an optimum heart rate band for the person using them, mostly that heart rate band is targetting fat burning, but elite sports people can use them to train for their particular event. If the actual heart rate goes outside the band, the watch or gym machine beeps, and the user can speed up or slow down to get back in the optimal range.
I have a little bit of a different use in mind, at least for the the early stages of the residency. I just want to track my heart beat, while sleeping, sitting at my desk, and the occasional walk to work.
Unfortunately it turns out the first heart monitor I purchased is only able to give me a summary of my heart rate over a period of time: the average and the peak, no where near detailed enough for my needs. This is despite me encouring the salesperson to show me some of his data, and being quite happy at the ten second coverage. As I’m hoping to just be able to upgrade the watch and use the same chest strap, I’m still wearing the chest strap to get used to using it during the night.
Live and learn!
A lot of products use accellerometers as a proxy for sleep it’s cheap and can provide good data. For me though, I wanted to monitor my brain and the Zeo sleep tracker is basically the only one of its kind.
The Zeo is a headband with three pads on it. These pads are woven with metal, which makes good electrical contact with the skin without being as annoying to place as the electrodes during my sleep study. The Zeo is not a medical device and won’t give as accurate results as a clinical sleep study, but it is very good for what it does.
The Zeo divides sleep into a number of types: awake, REM, light and deep sleep. The headband uploads its data over Bluetooth to my mobile and from the mobile up to the Zeo website.
It’s pretty easy to set up and use, but I’ve had great difficulty using it during Brisbane summers on those hot sultry nights every extraneous bit of clothing just increases the discomfort.
The Zeo website makes it trivial to download all of the data in a simple, detailed format, which is perfect for my use.
Probably the most important, and yet the lowest tech gadget that I’ve got is the humble tape measure.
The obvious things to measure is my girth, particularly my pot belly, but I’ll also be taking some measurements of my chest, hips, legs and arms for completeness.
Intra-organ fat kills, and on males, that fat concentrates around the stomach region, making it relatively easy to measure. I’m going to be measuring my non favoured leg and arm, assuming that I’ll have less muscle there and more fat.
I think it’s important to get consistency with the measurements, which is best done by measuring over particular anatomical landmarks, and measuring frequently. I don’t need to measure it every day, but I figure if I get into a routine with it, it won’t matter if I miss a few days here and there.
I’m using a Google form to submit the data, which means I can use a tablet, laptop or phone to submit the data wherever I am. Once submitted, the data goes into a Google spreadsheet. From there, the data can be easily exported as CSV and imported into my overall dataset.
The not-so-humble scales.
Time used to be where a scale measured your weight, and that was it. Now, you can get scales that will approximate your bodies fat percentage as well, it’s enough to make you want twice as much comfort food.
To get even fancier, the FitBit Aria scale I went with even uploads these measurements over wireless! All of the family can add accounts, and the scales will play a guessing game as to which member of the family is standing on it.
Now, the accuracy of the fat percentage is really questionable. It sends a tiny electrical pulse up one leg and down the other, measuring the resistance in the body. Fat, muscle, bone and water all have different electrical resistances, so the scales take a bit of a punt and estimate how much fat you have in your body.
The real issue is that it’s not forcing the electricity pulse through your entire body; there is a set of scales that comes with, essentially, a set of handlebars. You pick these up and the scales can send the electricity pulse through your hands down to your feet, ensuring a much better measurement. Unfortunately these scales aren’t licensed in Australia. I’m not too concerned about this, as I’m not all that interested in the exact measurement, just that it’s consistent over time so that I can see a trend. Based just on the week’s data I have so far, it seems very consistent.
The basic job of the pedometer is to measure the amount and intensity of your walking or running; there are many on the market.
Older style pedometers had a little metal ball bearing that bounced up and down with your hip movement, every time the ball bearing was forced upwards it would hit a couple of contacts, complete an electrical circuit, and another step would be tallied.
Newer pedometers use electronic accelerometers, which use fine strands of wires swinging in the breeze. On movement, these tip of these strands tends to stay still due to their inertia, but the root of the strand has to move, this forces the hair to bend, which changes its electrical properties and that is what is measured. Think of it like someone with long hair doing some headbanging, the tip of the hair and the root of the hair are moving at different speeds, bending the hair. The mechanism is actually quite similar to how our inner ear works out balance, except the inertia is provided by a small amount of liquid inside a tube. Inside the tube, there are very fine hairs that can feel the liquid moving over it.
Typically, these electronic accelerometers come in triplets, one to measure left to right, another for forward and back, another for up and down. These accelerometers are incredibly small, and are basically everywhere these days, including mobile phones.
The FitBit One is a pretty amazing bit of kit, it’s not much bigger than a couple of pieces of chewing gum, has a battery that lasts a week, stores minute by minute data and uploads it wirelessly over Bluetooth 4. Using the up-and-down accelerometer it even tries to work out when you’re going upstairs, which uses a lot more calories. It also has a sleep mode that I’m using, but I haven’t tried to correlate the data with the sleep monitor yet.
It hasn’t all been rosy with the FitBit though, more in another post on your data freedoms.