Thanksgiving Big Data and a Few Turkeys

If anyone has the scoop on what people are thankful for, it’s Facebook. They don’t need to tap anyone on the shoulder, send out any surveys or buy any data sets. The world’s leading social networking site is rich in “likes," data, and some of the world’s best data scientists.

And not just that, but Facebook also has enthusiastic members who thrive on holiday challenges like “write down 3 things you are thankful for every day for the next 5 days.”

To a data scientist an onslaught of information like this spells opportunity to discover what most of us, collectively, believe is good in our lives and in the world.

Sound like a project?

It certainly did to Winter Mason, Funda Kivran-Swaine, Moira Burke and Lada Adamic, a team of data scientists at Facebook.

So it goes to follow that, after anonymizing aggregate data generated by English speakers in the United States, they looked at status updates that contained “grateful” or “thankful,” as well as the word “day” preceded or followed by a number. They then aggregated and processed status updates that met the criteria through a text-clustering algorithm in order to identify what people were grateful for.

Some of their discoveries were to be expected; Family, friends, children, health, “roof over my head” and being employed topped the list. Interestingly, music hit the top 10, but pets did not. And through a finer lens, daughters ranked higher than sons, and “fur babies” ranked just under real babies.

The Deep and The Shallow

But take out the big stuff and what comes after, especially on a state by state basis, is a real eye opener. 

While people in recently drought-stricken areas like Texas and New Mexico expressed that they were grateful for rain, their counterparts in Virginia and California were thankful for YouTube.

Is the latter group shallow? Probably not when compared to Vermonters, who revealed that they had an obsession for Pinterest, or New Hampsharites, who showed their love for Netflix. And what might folks from Yoga-loving Oregon or Washingtonians, whose pick was “the ability to love,” think about Michaganers who are thankful for electricity?  Perhaps this map from will help find the answer.




Can Turkeys Know BI from Predictive Analytics?

We think turkeys better know because their lives could depend on it. Bruno Delahaye, SAP’s VP of Predictive Analytics Sales, says that turkeys who look at their world through a BI lens assume that they’re well loved.

“Day after day their owners are taking better and better care of them, ensuring that they’re very well fed and that they grow in the best possible conditions,” he writes. “And each day that the owners take a little better care of them, they really look forward to the next day. That is, until the Thanksgiving tragedy happens -- they become the most popular meal of the season!”

The same turkeys who use a predictive analytics lens might see being well fed and cared for very differently. Instead of focusing on past events, they might look at several years of data “to identify trends and patterns in order to gain a good understanding of what is coming next,” writes Delahaye.

It’s important to ensure that we learn from past events, analyze data and ensure that we anticipate problems in order to find the right strategy moving forward, he says. And though Delahaye's not quite sure about what turkeys can do to avoid their fate, he does provide a nice way to understand the difference between BI and predictive analytics.


Where Does Your Turkey Come From?

According to DataHero, a service that helps you visualize your analytics, Minnesota produces 45 million turkeys. North Carolina comes in second.


Need any more trivia, check out their blog where you’ll find trivia about this weekend's football games too.

Title image by foodllama (Flickr) via a CC BY-2.0 license.