Sentiment Analysis and Social Media

Reading Time: 2 minutes

 

I have two opinions about sentiment analysis technology. The first one is that social media have already suffered greatly through the actions of social media marketers, PR professionals and “internet famous” individuals. This technology could lead to broadcasting without a human bothering to listen as the human is replaced by algorithms and data analysis. This will benefit data analysts.

I do feel encouraged by another aspect. Automation. Social networks such as Twitter, Facebook, Google Plus, Instagram, Flickr and others will be self monitoring. I understand “Self Monitoring” to use data analytics tools, with Artificial intelligence to look for keywords and key phrases to create a human readable graphic to understand how brand perception is evolving over a period of minutes, hours, weeks and months.

We know the release cycle for apple products, Sony Products, expectations for the Oculus Rift and music festivals as examples. Twitter and Facebook have almost a decade of data about regional events, product releases and user reactions. If that data is processed then it could graph attention, interest and sentiment.

Current affairs affect how we perceive brands and products. If a social network was paired with sentiment analysis and current affairs imagine how it could react to the FBI V Apple debate. In the timelines of privacy enthusiast Apple products could be promoted as sentiment towards that brand progresses. In the case where people side with the FBI posts promoting other brands of mobile phones could be promoted in their place.

I love social media as a peer to peer communication tool and I hope that Sentiment analysis software will shift the social media practitioners’ focus away from getting followers to promoting their brand or products to people who are most likely to be receptive to the displaying of related information.

Hypothetically a group of friends could join a new social network, the friends they communicate with most frequently are suggested as people to follow and interact with. At this point no broadcaster, no media company, no brand are visible to the user. The purpose is for conversation, not broadcast. As people converse the Sentiment engine would analyse the tone and topic of conversation and according to this suggest related products. As data is collected over weeks, months, and eventually years so the software provides relevant information automatically, without scheduling.

Facebook, twitter, Sports tracker, Suunto and Google have ten or more years of data of the sports I practiced and events I have participated in. With that data they can see how passions grew and declined and how social networks changed from location to location and job to job. In effect social networks “know” me.

Social networks could ban animated gifs from my timeline and sensationalist writing could be avoided. I would be interested in a social network where I choose my friends  and then big data and sentiment analysis offer me the most relevant content. It would provide further information when we make positive statements.

When I joined social networks a decade ago they were blank slates in alpha or beta phases of development and we shaped social networks according to our conversations. Ten years later when you join a social network you are flooded with so much information that it is overwhelming. As it is time-consuming to pick and choose you either ignore the options completely or you select all of them and get noise. With big data and Sentiment analysis your conversation would automatically affect what you see, when and how often. As soon as you start being negative about specific things they fade away.