MIS 2502: Sentiment Analysis
The following Sentiment Analysis write-up was completed as an optional assignment for MIS 2502: Data Analytics.
Sentiment analysis, also known as opinion mining, is a field of data science concerned with discerning positive, negative, or neutral attitudes behind text. Traditionally, sentiment analysis has been conducted with the ‘Bag of Words’ method, in which fragments of a text block are compared to a dictionary of emotional words and phrases, scored as either positive or negative accordingly, and totaled to produce a net sentiment value. However, as machine learning and artificial intelligence develop, a method known as natural language processing (NLP) has begun to emerge. NLP programs strive to distinguish emotional phrases without a dictionary by placing text into context, recognizing parts of speech, and analyzing grammatical patterns (Sentdex, 2015).
With the rise of social media and online communication comes an influx of unstructured data in forms such as tweets, emails, and Facebook statuses. Sentiment analysis takes this messy data and converts it to a usable, quantitative measure, allowing organizations to gain valuable insight into their reputation, gauge the success of products, and identify common roadblocks stopping potential customers from purchasing a product or service.
The principles of unstructured data covered in MIS 2502 are closely aligned with sentiment analysis. The text data used in opinion mining, scraped from sources such as Twitter and Reddit, has no data model or predefined organization. In addition, unstructured text is omnipresent online, which reinforces the fact that 70% to 80% of an organization’s data is in unstructured forms, which was taught in the course.
The American Cancer Society (ACS) has implemented sentiment analysis to improve marketing for its Relay for Life fundraising event. After studying social media posts, the ACS’s research team found immense positive sentiment surrounding two events, the Luminaria Ceremony and the cancer survivor victory lap. Using this information, ACS leadership decided to promote these two events to attract more Relay for Life participants going forward (Bort, 2012).
Works Cited:
Bort, J. (2012, December 4). How Starbucks And Other Companies Use Complex Math Algorithms To Read Your Feelings Online. Retrieved April 25, 2019, from businessinsider.com: https://www.businessinsider.com/twitter-facebook-monitoring-2012-11
Sentdex. (2015). Sentiment Analysis – What is it? Retrieved April 25, 2019, from sentdex.com: http://sentdex.com/sentiment-analysis/