1
Twitter Sentiment Analysis: A Case Study for Apparel Brands
Abdur Rasool1,*, Ran Tao1, Kamyab Marjan1, Tayyab Naveed2
1School of Computer Science & Technology, Donghua University, Shanghai, China.
2College of Textile Engineering, Donghua University, Shanghai 201620, China.
*Corresponding author e-mail: 317089@mail.dhu.edu.cn
Abstract. Social media especially Twitter is providing a space for expression and
opinions, where users discuss various events, services, and brands. Entrepreneurs are
in continuous need of the feedback about their services to improve the quality and
quantity. However, due to the bulk amount of data, it’s difficult to detect the
consumer’s opinions. This article deliberates the problems about the Twitter data for
the sentiment analysis. Furthermore, it implements the text mining and document-
based sentiment on the preprocessed Twitter data through the machine learning
techniques, Naïve Bayes and lexicon dictionary. Our case study is to find the public
opinion about the top two apparel international brands and compare the positive and
negative attitude of common users about each brand. We found that positive reviews
of Adidas are more than the Nike while there is the slight difference in negative
reviews. It founds that people want to discuss the other brands as comparisons when
they are talking about just one brand.
1. Introduction
In the current century, the world is a global village due to the internet availability. According to [1]
54.4% of the world population was using the internet until Dec 2017. Now a day’s social media
(Facebook, YouTube, Twitter, and, content sharing sites, etc.) has efficiently and effectively shared
useful information. In the learning resources, it is statistically [2] verified that 71% of the internet has
been used through social media by the consumers. Studies [3] show that more than half customers
prefer to read the other’s comments about that products before purchasing. Thus public opinions are
the best source of feedback for business stakeholders about their products and services which enables
them to redesign the quality factor and disclose the opportunity of a new business [4]. The social
network like Twitter and Facebook provided the important marketing, selling, branding, and
promotional chances to the brands [5].
In the last decades, a hot research gets popular among the researchers to get the useful information
from all these social media especially from Twitter. It’s a mathematically procedural study of people’
thoughts and opinions which can be positive or negative about any product or event through the
natural language processing namely as ‘sentiment analysis’. Sentiment analysis is correlated with text
mining or data mining. The basic purpose of sentiment analysis is to assure the polarity of natural
language by performing supervised and/or unsupervised classification. Recently available sentiment
analysis techniques are useful for [6] political predictions, marketing strategy, e-commerce, and brand
reputation management.
In one study, Sharma et al. have performed an apparel brand study, in which he found the trust
level using regression tool for the Facebook data [5] but this study implements the sentiment for
