To find out, Facebook's computers analyzed approximately 1 million status updates from U.S. English speakers. It then broke down those updates based on context, demographics and content to figure out just what its users are buzzing about.
The social network first analyzed the correlation between the use of specific words and age. Unsurprisingly, they found that there's a positive correlation between age and religious words, family, and positive emotions, while there was a negative correlation between sex, sleeping, first person pronouns, school, swear words and negative emotions. In other words, the younger you are, the angrier you are.
Facebook also analyzed the correlation between word choice and friend count. While the correlation is not as strong as age's correlation to word choice, Facebook's data team did find that there's a positive correlation between friend count and second person pronouns, total word count, communication, religion, swear words and sex. On the other end of the spectrum, there's a negative correlation between friend count and past and present tense verbs, family and emotions.
Another interesting finding is the impact of the time of day on word choice. Not only do Facebook's users tend to talk about sleeping the most at around 4 a.m. ET, but they tend to talk about their work right before they head into the office. Positive and negative emotions are also affected by the time of day; Negative emotions tend to peak at around 1 a.m. ET, while positive emotions tend to peak at about 7 a.m. ET. More importantly, negative emotions tend to increase as the day progresses at the expense of positive emotions.
What type of status updates tend to get the most attention? Facebook dug into this data as well and found that positive status updates tend to get more "likes," while negative status updates tend to get more comments. Check out this image for the breakdown:
What do you think of the data Facebook has collected? Do any of these correlations surprise you? We want to hear your thoughts in the comments.