E-reputation prediction based on sentiment analysis Case study: McDonald’s

Abstract

In a competitive world, companies are looking to gain a positive reputation through their clients. Electronic reputation is part of this reputation mainly in social networks, where everyone is free to express their opinion. For this reason, this research is done to figure out an efficient way for the companies in order to keep a record of their performance by collecting data from Twitter. Twitter is considered to be one of the most used social media platforms, Natural Language processing techniques, in particular, sentiment analysis and text mining were applied on cleaned data scraped from Twitter. In this project we were able to discover absorbing results about our case study McDonald’s. This company was chosen for several reasons such as, availability of its data, its global popularity and the diversity of opinions about it. The results were interesting as some aspects have negative reputation while others have positive. Also, the performance could be observed over time. Although the insights could be very useful for firms, there are some limitations to the used methods such as, the ambiguity of natural language and sentiment analyzers could face difficulty in recognizing things like sarcasm and irony.