1. NUTRITION, QUANTIFIED SELF, AND MACHINE LEARNING

  2. by Georges Duverger
  3. I have been tracking what I eat since September 2013. 22 months and counting. More than 3,000 entries. 230 espressos. Throughout that process, I learned a ton about what works and what doesn't when it comes to logging food.
  4. I didn't realize when I started to record my eating habits that I really wasn't an isolated case. Pew Research Center conducted a public opinion polling in 2013 entitled Tracking for Health. They published some eye-opening findings:
  5. 60% of U.S. adults say they track their weight, diet, or exercise routine.
  6. 49% of trackers say they keep track of progress “in their heads.”
  7. 34% say they track the data on paper, like in a notebook or journal.
  8. 21% say they use some form of technology to track their health data.
  9. Of those 21%:
  10. 8% of trackers use a medical device, like a glucose meter.
  11. 7% use an app or other tool on their mobile phone or device.
  12. 5% use a spreadsheet.
  13. 1% use a website or other online tool.
  14. I was taken aback by how many people were still using such old-fashioned techniques (“in their heads”, really?!). I tried the most popular apps to find out why they weren't more successful and what they were missing. I quickly realized that they require a lot of dedication for very little return. It wasn't going to work for me, and apparently not for others either.
  15. The existing health and fitness apps prematurely optimize accuracy at the expense of usability and ease of use. This causes a lot of casual trackers to give up too quickly. It is a shame, considering that “88% of MyFitnessPal users who log their food for more than 7 days in a row ultimately lose weight” (source) and the act of “keeping a food diary can double a person's weight loss” (source).
  16. As a software engineer focusing on user experience for the last 8 years, I saw a huge opportunity to design a better solution to an extremely common need. A few failed experiments later, I ended up with a sustainable method that I have now been using for almost 2 years. The gist of it is a low friction entry mechanism combined with machine learning algorithms to provide daily nutritional insights.
  17. I recently left my job at eBay and I am now working full-time on this project—I have become the mayor of the Brooklyn Public Library in the process. I am a few weeks away from releasing a beta version. In parallel, I am assembling a team of health-conscious engineers, business folks, and nutritionists to help me shape this vision of better living through data.
  18. I would love to hear what you think! Please, send me an email and let me know.

  19. Made in Arlington, MA. Generated with Ivy. Styled with Backslash. All emojis designed by OpenMoji – the open-source emoji and icon project. License: CC BY-SA 4.0