This was my very first data analysis Capstone project, which was part of Springboard‘s three-month Introduction to Data Science course, which I completed in Q1 of 2018.
At the time it took me around two to three weeks to complete. Today, a project of similar scope can be done in under a week. If I were to repeat this project today, I would filter out the tweets by vendors selling COSRX products and include only customer sentiments. There are also better tools for visualization such as Tableau and Wordclouds.com.
I extracted tweets about cult-fave Korean beauty brand, COSRX and performed sentiment analysis to figure out what people were saying about the brand and its products. This kind of project is very useful for any business/entity that wishes to mine what customers are saying about them online.
The results in a nutshell: You are far from alone when it comes to those stubborn, ugly zits 😘
R libraries were used for tweet extraction, data cleaning, exploration, visualization, and topic modelling (LDA). For further reference, you may check out the project report and R code on my Github page.