Big Data Highlights Broad Human Behavior Patterns in New Study

Researchers from the ThinkBIG project at the University of Bristol have concluded that, by looking at patterns in huge data sets, including newspaper content, individual Twitter feeds and Wikipedia habits, it’s possible to spot things about collective society that might not be detected otherwise.

Its a remarkable revelation, made possible by the projects leader, Nello Cristianini a Professor of Artificial Intelligence and other scientists. During the study, they used big data to study the digital habits of large numbers of people to draw conclusions about how people behave, and to find notable similarities across demographics.

Trends in Newspaper Content Across Time

Researchers carried out two different studies before asserting their recent findings. The first involved looking at newspapers from the United States and the United Kingdom, published between 1836 and 1922. They found that the ways in which people devoted themselves to either work or pleasurable activities was largely dependent on seasons and weather patterns. As you might imagine, the word “picnic” showed up particularly often in newspapers during the summertime.

However, the printed content didn’t just mention how people spent their days. It also showed peak trends in mentions of fruits and other foods, which affected the diets of individuals from the past. Since diet often affects health, it’s not surprising that scientists also noticed repeating patterns about diseases. For example, newsworthy discussions of measles always reached a peak in late March to early April.

Although these scientists pored over data to check for health trends in the past, today’s healthcare providers are also increasingly turning to data. In the medical industry, they use data portals and secure storage methods to provide faster treatment and to facilitate the authorized sharing of patient information with fellow medical professionals. The convenience and security of electronic health records is arguably causing positive changes for providers and patients alike.

Decades from now, scientists may look back at how healthcare recordkeeping evolved, potentially using that data to draw their own conclusions about trends from today. For example, they might find that the enhanced collaboration between healthcare providers, made possible through electronic health records, allowed large percentages of patients to receive diagnoses quicker, which improved overall quality of life and facilitated faster treatments.

Tapping Into Social Media Use and Wikipedia Visits

The second study carried out by University of Bristol scientists focused only on the United Kingdom across a four-year span. It looked at collective sentiments from Twitter feeds, plus Wikipedia access habits. Data showed that, in the same way seasons affected what people did, they also potentially impacted mental health and how people expressed themselves.

Specifically, negative feelings were most commonly showcased across Twitter in the winter, and people were most likely to be anxious and angry between April and September.

A Different Take on Twitter Content and How It Relates to Mental Health

The University of Bristol scientists haven’t been the only researchers to look at possible connections between mental health and behavior on Twitter. A study published in the Journal of the American Medical Informatics Association analyzed 176 million tweets and tried to make sense of certain upticks in interest about mental health.

For example, there were spikes in tweets about mental health that correlated with national and international events, such as World Suicide Prevention Day. Robin Williams’ death by suicide also caused an increase in related Twitter-based discussions, which was unexpected.

In the above study, scientists discovered that, when mental health organizations tweeted to announce things associated with planned campaigns, the increased interest level within the public lasted for less than two days. In contrast, conversations about Williams death were higher than average across Twitter for 38 days.

Theoretically, scientists could also peruse data to find if certain times of the year, national tragedies or economic downturns made people more likely to be suicidal. If such links were found, perhaps mental health professionals could work especially hard to reach out and intervene through social media and other public realms to reduce previously identified trends that tie in with mental health crises.

Made Possible by Big Data

The research team from the University of Bristol was particularly excited about the opportunities made available to them courtesy of big data. Behavioral patterns uncovered during the course of the studies were long suspected, but were until recently nearly impossible to confirm. Scientists were only able to do so after looking at huge amounts of data collected across long timespans, and by viewing large amounts simultaneously.

These investigations resoundingly prove how useful big data can be in revealing repetitive and potentially damaging behaviors in society, and learning valuable things from such trends. What will another few years of research reveal about us?