How Twitter Can Make Us Healthier

It's easy to hate Twitter. It's not really a place for legitimate criticism or conversation due to its strict 140 character structure, and the ease of publication gives people who maybe shouldn't always be publishing every one of their thoughts a platform open to them 24/7. But Twitter does have plenty of worthwhile uses.

Not only is it a great method to deliver boots-on-the-ground news from around the world (the goings-on in Ferguson, Missouri being the most recent example), but the vast amount of information also allows for the collection of some fascinating data. For instance, what foods people are tweeting about.

Last week, the University of Arizona released a study in which they tracked America's relationship with food through tweets. Researchers analyzed 3.5 million tweets that contained food-based hashtags that were sent between October 2013 and May 2014. They then cross-referenced those tweets with the city they came from, allowing them to see which foods were being discussed most often in which states.

Now, plenty of the results were obvious. Texas, it's no surprise, has a lot to say about "brisket." And that predictability extends to our own state. Here are California's most popular food-based hashtags:

They were even able to break down the tweeted terms with more geographic specificity: Los Angeles was tweeting about #foodie, #tommynolans, and #tolucalake, while San Francisco was into #vegetarian, #dinner, and #brunch. Now, this is fun information and all, but what can we actually do with it?

The biggest issue is that there may be a hesitancy to take this information as valid because of how it was collected. That is to say, the study is fun and everything, but tweets are not exactly an unbiased survey. People tweet more often about exceptional experiences they have had than daily mundane occurrences, and it's the latter that are more valuable in terms of information collection. Also, Twitter includes more urban people than the general U.S. population, so by only looking at tweets to come to conclusion, a large portion of citizens are ignored.

So, what's the point in collecting this data? Is it just cute to see everyone's differences?

Not exactly. Surprisingly, there may actually be some legitimacy behind these numbers. To point this out, Mihai Surdeanu, one of the researchers behind the study, told me about the interesting correlation between food that was being tweeted about and immigration patterns.

"We discovered that there is a high concentration of Vietnamese food on the West Coast (not surprising), but also in Houston, Texas, which was surprising until we found out that there is a major Vietnamese community there," Surdeanu told me in an email. "Similar patterns were discovered for other diets (Italian, Mexican, Japanese, etc.). The fact that these trends match what we know about the U.S. suggest that this data (even if biased) on average captures typical diets seen in the U.S."

When you can capture accurate information about the average diet, you can begin finding food-based correlations for everything from political parties to health problems:

What can this information lead to? For starters, knowing what areas are most likely to have a high incidence of diabetes can help funnel money in the right direction.

"Improving a community's public health (e.g., by improving diabetes awareness in communities at higher risk of diabetes), or nudging individuals that the model detects to be at risk of diabetes towards a better life style," said Surdeanu. "This is where our future work is going."

Who would have thought that Twitter could actually change our health for the better?