Artificial intelligence analysis of tweets predicts people’s moods
Do you have dark thoughts in the middle of the night but wake up motivated as if nothing happened? A new artificial analysis (AI) of tweets worldwide confirmed that you’re not alone. An enormous analysis of over seven billion words from 800 million anonymous tweets exposed collective tendencies in how manners of thinking tend to oscillate over the progression of each full day. The research article, “Diurnal Variations of Psychometric Indicators in Twitter Content,” was published in the journal PLOS ONE.
For this groundbreaking study, researchers from the University of Bristol in the UK used artificial AI to follow specific word usage on Twitter from a sizeable population sample during many full-day cycles. They examined tweets over the course of four years apart from tweets that were sent during big holidays due to the skewed nature of language introduced by each specific holiday.
Interestingly, they identified that the first peak communication period on Twitter began approximately from 5 a.m. to 6 a.m. These initial tweets of the day had language that was mostly linked with motivation (e.g., ambition, influence, and success) as well as personal trepidations about the upcoming day.
Also, the researchers found that the language used in these early morning tweets contained words associated with critical thinking.
Erratic Thinking Before Dusk
However, when the researchers examined twitter language from just a couple hours earlier from 3 a.m. to 4 a.m. they found that words linked with a rash, emotional, and existential way of thinking.
In a complete flip of the motived morning mulling, these tweets were also correlated with the expression of negative emotions.
Tweets Spot Global Communication Trends
Although the results of this mega AI analysis of tweets are not entirely surprising, it shows that huge amounts of anonymous data, whether from tweets or another source of communication, can spot global tendencies in the way people think and feel throughout the day.
These findings suggest that the changing way someone may think and feel throughout the day can be generally predicted based off the way they communicate.