Sarcasm Presents Data Problem For 2016 Presidential Candidates

Big data has long been used by firms to gather information on potential customers and target specific subsections of the population. The advent of social media has made data collection even more useful for firms who are looking to get an idea of the public's perception of their brand or the success of a specific campaign.

Next year's presidential race is expected to similarly use social media metrics to measure the impact each candidate is having on the public; however, data analytics firms say measuring public perceptions is becoming increasingly difficult.

Related Link: New Book Reveals The Repercussions Associated With The Global Economy

Sarcasm Detection

Because much of what is posted, shared or retweeted on social media has a humorous element to it, the messages can be cryptic for data analytics software. Sarcasm, in particular, has proven difficult to identify when it comes to data analysis. Phrases like "sure it is" and "I'm shocked" can have two very different meanings depending on context, but computer algorithms are unable to detect which expression is intended.

Predictive analysis firm Haystaq found that certain word combinations can be taught to a computer to help weed out sarcastic tweets. For example, the firm found that tweets using the word "classy" were typically expressing a positive view. However, when "classy" is combined with "Trump", the majority are painting presidential hopeful Donald Trump in a negative light. The phrase "stay classy" almost always indicated that the tweet was written to mock the subject.

Advertising Dilemma

Data firms are working to sure up their analyzation techniques this year in order to improve their accuracy when the race for the White House heats up in 2016. Candidates will want to target supporters with online advertising aimed at relevant online groups and inaccurate data readings could create problems.