Julia Silge teaches sentiment analysis in R using tidy data principles. The course uses four real datasets: geocoded tweets, Shakespeare plays, cable news speech, and pop song lyrics. You'll learn three types of sentiment lexicons that score words for positive/negative attitudes, numeric polarity, or emotions.
Chapters
Hi, I'm Julia Silge and I am excited to welcome you to this course. Sentiment Analysis the Tidy Way Text data is being generated all the time in the world around us, and sentiment analysis is a way to measure the attitudes and opinions expressed in text. In this course you are going to approach sentiment analysis using a specific idiom and approach, and that approach is using Tidy Data Principles. If you've used dplyr for data manipulation or ggplot2 for visualization, then you have used Tidy Data Principles. And here you are going to apply those same tools and ways of thinking and working to text mining and sentiment analysis. We're going to talk about the sentiment lexicons that are used in sentiment analysis and you are going to apply these ideas and knowledge to real text data from the real world in form Case Studies in our Chapters Our four chapters in this course will use four data sets. First, a data set of tweets that have been geocoded so we can see what kinds of words Twitter users in different locations use. Second, a collection of comedies and tragedies by Shakespeare. Third, a data set of speech from cable TV news shows and finally, pop lyrics from songs over the last several decades. These case studies demonstrate how you can use these powerful techniques with diverse kinds of text and will allow you to practice your data handling skills and learn about the ways sentiment analysis can be applied. You'll see how you can get to real insights from raw text data. We're going to start by exploring some sentiment lexicons. These lexicons are lists of words that are scored in some way according to the emotion or opinion content of the words. The words can be scored, like this lexicon, in a binary fashion, either positive or negative. Or they might be scored like this lexicon with a numerical score that runs from negative to positive to indicate how negative or positive the word is. There are three lexicons that we will explore during this course, and the last one, shown here, categorizes words into specific emotions as well as positive and negative. Lexicons like these don't contain all English words because many words are pretty neutral. The choice about which lexicon to use in analysis often depends on what question you're trying to answer. Okay, now it's time for you to explore the lexicons yourself to get you started on the path of sentiment analysis using tidy data principles.
