Brand Analysis in Social Networks Using Deep Learning Techniques

Abstract

In recent years, the importance of social media data has increased with the developments in information and communication technologies, and data volume, velocity, variety, veracity, and value have been affected by these developments. Because of the popularity of social networks, the analysis of social media data has also become an important issue for large companies whose brand identity is very crucial. User comments, shares, and explanations in social networks can be used to obtain information about the brand and product. Besides, deep learning techniques, which have become popular recently and provide high accuracy, can be employed for big data analysis in social networks. The number of studies examining the brand image in social networks is quite limited. In this context, we developed a model that performs brand analysis using deep learning techniques in social networks by considering the Starbucks Coffee Company, one of the world's largest coffeehouse chains. We trained our model with Faster Region-based Convolutional Neural Network (Faster R-CNN), Single Shot Multibox Detector (SSD), Mask R-CNN, and You Only Look Once (YOLO) algorithms. We then tested the model on data from Instagram and compared the results. In the light of our results, we have shown that analyzes using deep learning techniques in social networks can significantly affect the image of companies and their brands.