Sentiment tracking is a technique in social media analysis used to determine the positivity of an online comment or tweet. Using various algorithms that detect the words used determines whether the tweet or online comment is negative or positive.

It started around 2010 using word lists to determine how positive the results would be. This was a very basic principal that took words and gave them a ranking of sentimental value. For example swear words would be low on the negative scale whereas words associated with happiness, success and achievement would be ranked higher and more positive. An obvious floor in this technique is that it meant that campaigns could appear to be both more successful and less successful due to gaps and its unreliability to detect the sentimental value of a sentence, not just the words.

Initial techniques of sentimental tracking were unreliable, overtime it developed using newer technologies. Social Media Analysis' needed a way to detect whether words were positive or negative. An example of this is “Sick” which can be used as both positive and negative.

DataSift, a social media analyst claims to have the highest reliability rate of sentimental tracking with 70%.

“There has been a clear shift in the last three years — the difficulty with sentiment analysis really is about understanding the context of it, and the tech definitely has got better. We’re starting to bridge the gap, and we’re way beyond word lists now”, said Halstead.

This is often used to track online campaigns to see the outcome of how successful it was and what were the reactions of those that took part of were involved in the campaign.

John Lewis’ Monty The Penguin Christmas campaign towards the end of 2014 used sediment tracking in it’s social media analysis to evaluate how the campaign was perceived and how its customers responded to the campaign. This allowed the creators of the campaign to use to see how they can improve this and lead to an even more successful campaign.

The sediment tracking for the campaign revealed that 97.7% felt Positive/neutral about the campaign, with only 2.3% feeling negative.

I also ran some of my own tests using an online service iFeel 2.0. This uses 20 different methods of sentimental tracking and displays them visually. I ran two simple tests. One was the sentance “I love You”, which i obviously predicted to be positive and the other was “I Hate You” which i assumed would be negative. The results were accurate to my predictions and displayed as shown:

Tom.