Can A Picture Be Worth A Thousand Numbers

My name is Laila, and my little experiment is about 3D body models, which is a relatively new way of representing and tracking the human body.

But first as a quick background, I’ve been tracking my weight for a few years, and more recently have been tracking other people’s weight and physical variation, with a company that I co-founded called Fitnescity.

So when you think about weight as we all know it’s basically a number, and that gives you one specific perspective on your body. For example these people and I have the same weight and height, but the one on the left has higher health risks because of body shape. It’s also not that easy to differentiate between good and bad weight.

So for example, when my weight increases I always try and convince myself that I gained muscle even if it’s not true. Because it’s a bit hard to get an accurate measurement of body fat versus lean mass at least not in your home on a daily basis.

And so about a year ago I heard someone say that in the future, everyone will probably have a digital body model as part of their digital identity. And they would use that to buy clothes online or even track their health and fitness. And I found that quite intriguing actually because by scanning something which is how you would get that model, you basically get a dimensionally accurate replica of that object. So it’s a little bit like your accurate virtual self that you’re seeing and tracking.

And so I was curious about what I would learn from it from a physiological perspective obviously, but also may be from a psychological behavioral one. And so what I did is I used a 3-D body scanner, which is basically a machine that uses multiple cameras to capture information. And once the data is processed it can be visualized as a 3-D image.

And so I did this multiple times over a year, and during that time I continued to track with everything else that I was tracking. And so my first scan was in July of last year, and it was actually here in this location but it was in a different one, where I wasn’t given this scan data but just the end user access to the model. And the result was something like this, and I thought no, it’s cool, it’s okay but I also thought that there is probably a little bit more that could be done with this scan data.

And so starting from here, August of last year I built my own tool with the help of a small team, which means I was also able to add my body model to the same platform that I have all my data which is nice. And so at this point I only had one scan, so I wasn’t able to compare anything yet but what I did first is I started to play a little bit with this scan and measurement data. So one of the first things that I did, I learned to build a target volume model, kind of like you would set a target weight goal. So this is sort of like a futurist self.

The second thing I did is actually use the body model as an interactive visualization platform for other tracking data. Because I found that for some data it’s much easier to view it directly on the body. For example, if you track body fat locally you can just visualize the value interactively on the body. And then the coloring code can show you easily, for example whether the change was positive or negative so hopefully a visualization benefit.

And so at this point I finally got my second scan, so I was able to compare. So the first thing I did I compared the models visually obviously, but also compared them mathematically. So compared them by matching this scan data and actually computing the change, positive or negative locally by body parts.

And so result we see here for example the belt is shown in green when there is weight loss or when the volume decreases, and in orange when the opposite happens. And there’s also this sort of like heat map which shows you how the body is changing with you know coloring code that indicates the type of change. And the way it is done is actually by measuring the volume of the body, the volume in liters locally for each small body part.

And so this actually led me to another second stage in my exploration where I started to use the body model as a source of data itself, which was initially my intention. So for example, I extracted body volume for instance. I had no idea what my volume was in liters as a human being. So it turns out that this was the trend. And then what I did at this point is I also compare the out of curiosity with my weight soon you will see in kilograms.

Then interestingly, they didn’t follow the same trend and that has to do with our density, so human density being weight divided by the volume. So our density as humans changes over time because our body composition changes. So for example here, my density increased which means my body fat actually increased a little bit.

And so connecting this with what I was saying at the very beginning, so body fat calculations for a body impedance scale which is what I use on a daily basis have an 8 to 10% uncertainty. As opposed to here, it’s 1 to 2% if you extract body volume, which means that it’s you know kind of hard to argue that you gained muscle.

So another thing I did also with this is posture analysis, so I found that it was actually useful for analyzing posture, so I found out that the head has a small misalignment, which is for a while I was getting frequent injuries with exercise.

And so, what did I learn from this experience in summary? I did find that the body model was a good tracking and good visualization tool in different ways as I explained. But for me there was actually another hidden and kind of surprising element, which is that I found that probably the most compelling use case or the most compelling thing is more on the psychological behavioral part. So the way I like to explain it is think of it that you’re tracking your weight number or any other number that you kind of care about to a certain extent. And every time before measuring or looking at that number you get that feeling of you know, what I going to get. So this is a little bit the same, except that it’s very much amplified, probably because it’s not just a number that you are seeing but maybe yourself. So in terms of next, actually I decided to extend this experiment to a larger number of people, to understand these specific dimensions. So I want to understand for example whether visualizing your future self and your progress towards that, whether that’s something that could have some sort of impact on your behavior when it comes to dieting and exercise. So I’m doing that right now with Fitnescity, and we have about 100 people on the list and we are just actually opening it up to a larger number of people.

So if you are interested in the results, or if you are in New York some time and would like to try it as well let me know. My email is laila@fitnescity.com so if you have any comments, ideas or suggestions I would love to read them as well. And I’ll be here for office our, thank you.