Trying to predict weight loss using gut microbiomes

I think it’s really interesting that we can transplant poo from one person to another and solve all their problems. The idea itself seems so completely crazy, you can come up with a really clear picture of the first doctor to give it a go. But it works!

Behind the procedure is a lot of new science about gut microbiomes. We know that they have something to do with the amount of weight you’ll lose from a dietary change, but we’re not sure specifically what’s causing that. While there are studies which outline relevant biomarkers in broad strokes, we thought it was important to delve into a deeper justification of why.

Unfortunately, we totally failed. Machine learning has a few drawbacks with less-than-ideal data, and the weight loss dataset we were working with suffered from a lot of them. But while we weren’t able to get mindblowing results, I got a lot of experience with data analysis, data mining, and practically applying the machine learning techniques I already knew.

You can read the paper here