Module 4, Discussion 5: Analyzing Text Data

Overview

Module 4, Discussion 5: Analyzing Text Data

Introduction

In your initial thread, answer each of the following questions. Consider writing your responses in Word and then pasting them into your thread as to avoid the possibility of losing some of your work. Before you submit your initial or response posts, evaluate your work against the
Discussion RubricDownload Discussion Rubric
.

For this discussion, please follow the instructions below.

1. Think about some of your favorite celebrities (10 at least) and write them down.
2. Go to their Facebook pages. For each celebrity, choose three (3) posts.
3. Make a list of keywords that have a positive connotation or arouse positive emotions (e.g., love, excitement, etc.) Next, make a list of keywords that have a negative connotation or arouse negative emotions (e.g., hate, war, bias, etc.) Do NOT read the whole post yet.
4. Give each positive word a score of +1 and each negative word a score of –1.
5. Make an Excel spreadsheet and record the data. Use the ‘Sample Spreadsheet Download Sample Spreadsheet for Discussion 5’ as a roadmap.
6. Sum up the scores for each post. A positive overall score indicates a positive tone while a negative overall score indicates a negative tone.
7. Now, read the post very carefully. Are the results you got in line with the overall tone of the post after you read it?
8. Do you think algorithms like the one you used above are 100% accurate? Can these algorithms relace human interpretation of text?
9. Specifically, does the connotation of a word stands on itself or is context-dependent? Please cite at least two examples (e.g., I so hate your capriciousness in our intimate times, you devil! Despite the use of the word ‘hate’, this sentence carries positive emotions.)
10. Think of a sentence where the above algorithm returns a positive tone while the tone of the sentence is negative. Here is an example. 'You are the most wonderful, most perfect, most extraordinary, most out-of-this-earth, besteeesssst boyfriend of the century! Loser!' (The sentence is sarcastic, hence negative.)
11. How do you think we can improve such algorithms to detect the tone of a sentence on digital media?\
12. Submit a detailed report of all of the above steps through the discussion board.
13. Submit a copy of the completed spreadsheet with the assignment.