As we edge further into the 21st Century social media is generating more and more of our personal data. Data is collected on everything from our browsing habits, to our posts, likes and friends; images we like and share, to comments and opinions. The shear volume of data generated by social media is staggering. Every minute of every day:
- nearly 10,000 tweets are sent on Twitter
- 400 hours of video are uploaded to YouTube
- 216k photos are shared Facebook.
- 4 million photos are liked on Instagram
[source: DOMO – Data Never Sleeps 4.0]
And this volume is growing exponentially. 90% of the data in the world today has been generated in the last two years. This data is so important to advertisers that Facebook Inc. – arguably at the epicentre of consumer social data – is valued at more than $400 billion.
There are, however, other, deeper, applications of this social data, but they require the use of data science techniques to access the insight. Some businesses and groups have commissioned research on social media data, and applied machine learning algorithms, to uncover insights that are more socially (and ethically) responsible.
How Data Science Helps us to Understand Human Behaviour
Pre-empt Self-Harm
In one recent study, data scientists have demonstrated that Instagram can reveal if you are depressed based on your activity on the site. A new study published in EJP Data Science, finds depression reveals itself in your Instagram feed. The correlation was so strong that researchers were able to identify markets of depression in pictures posted before a person’s diagnosis.
Such research is not alone, and a similar trend was found in a youth social media study between the university of Cambridge and ULM university, where analysis was conducted on 32,000 pictures posted on social media depicting non-suicidal self injury (NSSI), and their corresponding comments. By applying data science techniques to social media usage, both the EJP and Cambridge studies have generated a staggering amount of insight into the personal traits of social media users at risk of NSSI. Such understanding could go a long way to helping those suffering from depression or other forms of mental illness online. The better we understand how depression manifests itself online the better we can support others who exhibit those online traits. Searching the hashtag #ritzen (cutting) on Instagram now prompts a pop up with links with information on getting professional help, along with a warning about how disturbing the images can get.
Sarcasm Detection By AI
The application of such data is not limited to simply understanding users either. Many have attempted to martial the power of data science in an attempt to improve our online experiences. One such attempt has come from MIT scientists who have developed a new artificial intelligence system that can detect sarcasm in tweets better than humans. A seemingly trivial application, but it’s actually an advance that may help social networks automatically identify hate speech and abusive comments, a technique that is extremely pertinent given the events last week in Charlottesville.
By using data science to real in user data, and combining it with artificial intelligence, MIT has begun to bridge the gap between recording user data and using user data. Detecting the sentiment of social media posts enables companies to track attitudes towards brands and products, and identify signals that might indicate trends in financial markets.
New Frontiers
In terms of the usage of data science, there is undoubtedly a cause for good here. However, harvesting data is not without its complications. In the UK the government is looking to implement a new data protection bill that will give people more control over their data. This means that businesses will require much more consent for its use, and force social media companies and online traders to delete their personal data; and with the European General Data Protection Regulation legislation looming, businesses across the world are going to have to be extremely careful with how they gather and use consumer data to generate insights into human behaviour.
Despite an increase in legislation surrounding data science, the future remains bright. So long as organisations structure their data collection practices within the confines of the law, we will be able to continue our attempts to understand online human behaviour. Just as psychologists study the way we live, think and feel, data scientists are studying the way we carry ourselves online, and how that relates to the complexities of behaviour. Data science could indeed open up the next frontier of psychological study.
