Case Study

Sentiment Analytics

Sentiment Analytics. Doing It Right.

The traditional sentiment analysis using Natural Language Processing (NLP) method has its limitations. We employ the Machine Learning Method that processes unique keywords including dialects and multi-language that exist in a single sentence and many other peculiar cases that cannot be processed by NLP.

Our level of sentiment accuracy is 79%. This figure is supported by years of scientific research (and white papers) in Machine Learning processing of a large volume of social data.

Compare sentiment between two or more products.
Sentiment analytics was deployed to compare sentiment for each users (marked by each coloured bubbles) to measure the reaction between two products. Initial findings, as seen below , revealed overall sentiment from different groups of users was skewed towards positive.
Measure public perception before a crisis happen.
Some crisis can be contained because it occurs within a similar interest group. Social analytics allows companies to predict the likelihood of an event (or campaign) to turn into one. We combine people, process and technology to obtain these insights from data science.


Social Data


unique users