CASE STUDY: ZimGo Social Politics
Sentiment Analysis has many practical applications to social studies. Politics, for example, provides a wonderful opportunity to showcase how powerful Sentiment Analysis can be when applied in context. Given the recent US Presidential Election, we thought it would be fun to put together a quick, real-time monitor using the ZimGo service to monitor the two top candidates- Donald Trump and Hillary Clinton. Given how much the media credited President Obama’s use of Twitter for his victory, we decided to use Twitter as our source for analysis. The system updates in real-time, but given the complexity of crunching big data, we decided to sample at intermittent points. For the purposes of this effort, a score of 50 represents neutral sentiment. The higher the score over 50, the more positive the sentiment.
First, the overall scoring,
It may be of surprising to some that Trump positively leads the sentiment with immigration; while others may not be surprised that people don’t think of him as the guy that will spread the wealth around. Interestingly, the overall score is a lot closer than you might have thought- they really were neck in neck:
The individual trends are significantly more revealing- there are some big swings particularly on the topics where they were both struggling. It could be interesting to track down what happened on those days to influence these major swings:
These charts demonstrate the tremendous value to anyone wishing to promote their candidate and strengthen those areas where sentiment is negative (under 50). News and polling organizations may also see the value to supplementing existing measurement tools with this real-time read on how the public feels as revealed by social media activity. Unlike a public poll, this is unsolicited and the tools are simply reading and scoring the posts.