farmdata

It’s Time to Eat

At Jawbone, our food-tracking features not only provide you with the most complete tools available for weight management, they also give our data science team an exclusive look into the eating habits of the UP community.
When diving into the millions of food choices logged around the world, we are able to extract detailed information on the foods eaten more often at a particular time of day. We use this information to make UP smarter and more intuitive, but we can also use it to tell interesting stories about the broader UP community. So, for the first time ever, we’ve created a ‘global menu’ of foods that people prefer* at certain times of day. Let’s take a look at the selection:


farmfresh_menu_shortened

 

Now let’s get into the specifics. Not surprisingly, the five beverages logged most frequently are water, coffee, soft drinks, green tea and beer (in that order). We see that our users prefer their beer at 1:00am and their coffee from 5:00am to 8:00am. On the flip side, people generally don’t like coffee at 1:00am, or beer in the morning. This insight helps us make smarter suggestions to our users—if it’s morning and you’re logging a meal, we’re not going to surface beer as an option for you to log at that time.

 

 

Even more insight can be gleaned from looking at groups of foods as they’re consumed over time. For example, notice how oils, fats and sugars are preferred in the wee hours of the morning. This fascinating data helps to power our Insight Engine™. After identifying that people tend to make less healthy choices at night, our Data Science team can generate Insights to help you eat better – and send them to you at just the right time.

 

 

And that’s just a taste of what we can do. Here’s an interactive visualization of the top 100 most-logged foods and beverages (excluding water) from the UP community. What insights can you glean?

 

Mouse over a line to see what food it is

 

Technical Notes:

* For each food, we count the number of times the food was logged relative to all other food (daily proportion). At each hour of the day we calculate the number of times a food was eaten relative to all other foods eaten at that hour (hourly proportion). ‘Preference’ is defined by comparing the hourly proportion to the daily proportion.

Data is presented in aggregate and taken from millions of food choices logged by members of the UP community globally.

 

About The Author

Karl Krehbiel

Karl is a data science intern for the summer of 2014. He is a rising senior studying statistics at Harvard and enjoys cool data, playing sports, FIFA video games, and the unlimited Nutella at Jawbone headquarters!