Is Beaumont, Texas the saddest city in America? Yes, if we’re to judge by Twitter prose, which is exactly what a team from the University of Vermont did. Their study found that tweets from Beaumont contained a whole lot of curse words and not enough “happy” words, such as rainbow, love, beauty, hope, wonderful, and wine.

In this new creative pursuit to quanitfy happiness, which The Atlantic broke down in layman’s terms, researchers analyzed 10 million geotagged tweets, creating a chart called a “hedonometer.” Each tweet was then coded for its level of happiness based on the frequency of words deemed “happy” or “sad” by Amazon Mechanical Turk (an internet marketplace wherein users post jobs that can only be completed by humans). Awesome, sleep, deal, Christmas, and lol all signified happiness in this study, while gone, no, boo, hurt, smokeugly, and a smattering of swear words were measured as indicators of sadness. Each city’s use of such words  was compared against the U.S. average:

The "hedonometer"

Of the 373 urban areas studied, Beaumont had the most negative words and least positive, making it the unhappiest city in the nation. Two other Texas spots, Texas City and Port Arthur, also landed on the top 15 Saddest Cities chart, in spots three and 13, respectively.

Though Beaumont’s “lol” usage was especially high, it apparently didn’t add enough joy to counteract the city’s ubiquitous cursing. The researchers took note of the large amount of swearing in certain cities, The Atlantic said, and are eager to dig deeper into the phenomenon, which they have labeled “geoprofanity.”

Where were the happiest spots in the country? Napa, California was the top city (presumably due to the frequency of the word “wine” in their tweets) and Hawaii was ranked the happiest state (no explanation needed).

Cities with more tweets overall were sadder, reported CNN. The Vermont study interpreted this to mean that higher levels of technology and connectedness may be correlated with unhappiness.

The researchers acknowledge that their study ignored context when ranking these keywords, but claim that with a sufficiently large dataset, it still offers “reliable results.” The Atlantic also noted that the study didn’t consider the unique ways that different communities and cultures might express happiness.

Despite scientific flaws, the research team says their results matched up with other measures of happiness. Gallup surveys, income, and obesity levels correlated with the Twitter assessment.