Saturday 20 August 2016

Figuring out Feelings

Understanding well-being and how we feel is very much in our consciousness. Organisations now see staff well-being and personal happiness as key components of a sustainable way of working and a big influence on staff retention. At a societal level we are concerned about how stressed and unhappy our youth are, how lonely our elderly might be and the increasing suicide rates in many developed countries. If we had a way of measuring and trending how people feel, that may go a long way in designing early intervention programs and reviewing the efficacy of the various initiatives deployed by organisations and society itself.     

I read something which got me thinking a little more about this. Researchers at the University of Manchester and Indiana University have put some science behind this idea of  ‘sentiment tracking’.

They looked at how global emotion and mood, as measured via something like Twitter, could predict stock market activity. They investigated whether measurements of collective mood states derived from large scale Twitter feeds, correlated to the value of the Dow Jones Industrial Average (DJIA) over time. They analyzed the text content of daily Twitter feeds by two mood tracking tools, OpinionFinder that measures positive vs. negative mood and Google-Profile of Mood States (GPOMS)

Their results indicated that the accuracy of DJIA predictions can be significantly improved by the inclusion of specific public mood dimensions. If you want to predict closing prices on the Dow Jones, have an eye on the Twitter feed.

This raises some interesting questions about how we look at social media data or activity. It may justify a more qualitative approachWhen it comes to social media marketing, we may need to look at how our followers feel, rather than just counting them.

While causation and correlation could be difficult to nail down, there is perhaps something in this for the Data Scientists among us. There is a plethora of online content generated by users across all forms of social media, text, images, video. 

Parsing online content and cross referencing this with behavior may create some predictive tools or highlight some change in public mood. Looking at which emoticons are used and how frequently may be one place to start. Too many ASAPs in emails might signal someone under a little stress, lots of work emails sent at night or weekends could be from someone on the road to burnout. 

The take away is that with a focus on the Internet of Things, we could be relying on devices to tell us more about the world and how we live our lives. However a lot of this information could already be out there, we just have to look for it.  


No comments:

Post a Comment