Researchers have been trying to study the biological rhythms of the human body for a long time, but have often been plagued with small, homogeneous samples. Researchers at Cornell University decided to use Twitter as a means to create a very large data sample; over 509 million messages. A key characteristic of this data is it comes from across the globe and not a single culture. What they found was not exactly what everyone expected. Using a lexicon to analyze the text for mood, the findings showed how time of day, work, sleep, and amount of daylight all impacted ones mood. The midday-down though, believed to be caused by exhaustion at work while still having hours to go, while found across the world, does not appear to be associated with work. While work can intensify the experience, it occurs on non-work days. Non-work days did show a higher average positive attitude.