Projects
2019 Internship Presentation
At the end of my 2019 internship with TJX, I compiled my work into a PowerPoint presentation and presented it to management.
Text Mining and Sentiment Analysis
One area of data analytics that has been growing in recent years is text mining. Text mining is the process of scanning through large amounts of unstructured data and transforming that data into structured data for analysis. Text mining allows businesses to gain important information that would have been lost inside a vast array of unorganized data. The unstructured data can come from a variety of sources such as websites, text documents, and social media. Text mining software searches the source and finds potential similarities through complex algorithms. It is different from a traditional keyboard search because the algorithm is capable of matching similar words, misspellings, and slang words. The data is quantified and can be integrated with traditional forms of data analytics, in the case of our MIS 2502 class we would use Excel to store the new data.
Another emerging area of data analytics is sentiment analysis. Sentiment analysis is very similar to text mining in that it uses computer algorithms to decipher large amounts of data to better understand it. Sentiment analysis is used to discover the emotional opinion of a text document. In the case of social media, sentiment analysis is possible for detecting the opinion of a user. For example, if a business is trying to garner public opinion on a topic, they can use sentiment analysis to search social media posts, such as tweets, to better understand customer’s opinions. Sentiment analysis can take the social media posts and classify them as either positive or negative. Since sentiment analysis is still a developing technology, developments are still being made to improve this process. One flaw in the technology is accurately detecting sarcasm to which the analysis may incorrectly mark a sarcastic post as positive when it is in fact a negative post.
These two emerging ideas further build on what we have done in MIS 2502 because the analysis can be done in RStudio. For example in the instance of sentiment analysis, a computer software can search through the document and quantify the similarities of words and ideas. The software will create an Excel worksheet which can be imported into RStudio and further analyzed through topics which we have covered in class such as decision trees to determine if a word is positive or not positive (negative). To find if words are similar in the case of text mining, using the clustering package of RStudio would be appropriate and helpful.
References
Bannister, Kristan. “Sentiment Analysis: How Does It Work? Why Should We Use It?” Brandwatch, 26 Jan. 2015, www.brandwatch.com/blog/understanding-sentiment-analysis/.
Rouse, Margaret. “Text Mining (Text Analysis).” TechTarget, Oct. 2013, searchbusinessanalytics.techtarget.com/definition/text-mining.
“Sentiment Analysis.” Sentiment Analysis | Lexalytics, www.lexalytics.com/technology/sentiment.
“What Is NLP Text Mining?” Linguamatics, 8 Feb. 2018, www.linguamatics.com/what-is-text-mining-nlp-machine-learning.
Sonny Shines Window
I created a web application capable of receiving orders from a form and sending them to a database on a server. The website also includes a login page which provides users with credentials to access the orders. The link to the site is http://misdemo.temple.edu/tug18271/SonnyShinesWindow/ and the login credentials are as follows (username: “admin” password: “clean”).