How Social Media Shapes Buying Behavior

Table of Links

1 Introduction

2 State of Art

3 Projects Selection and Data Collection

3.1 Blockchain data

3.2 Twitter data

4 Data Analysis and Results

4.1 Ethereum wallets and Twitter users and 4.2 The communities

4.3 Hashtags

4.4 The role of the social network community

5 Conclusions and References

2 STATE OF ART

Over the past decades, social media data have been studied to understand people [1, 3] and exploited to develop strategies for promotion, prediction and engagement in different !elds [8, 11, 25]. A growing body of existing research has analysed the dissemination of information online, investigating how buying products or joining communities is affected by the mechanics of social networks [20]. A considerable amount of work in this area has been done on sentiment analysis and opinion mining using Twitter. This is due to the fact that Twitter data is mostly composed of text, i.e. tweets. Twitter is largely popular for its institutional role of showing news and reports, but also renowned as the medium where users freely share their true feelings, write what they are doing and discuss a wide range of topics, which include places, people and products [11]. The purpose of sentiment analysis, thus, is to understand a user’s opinions and attitudes about different topics through the texts they have written [15]. Even if generally the understanding of communities in social network is performed through sentiment analysis and text mining [15], in this work we will focus on Social Network Analysis (SNA). In social networks, different mechanisms generate various structures such as friendship networks, mention networks, hashtag networks, etc. Generally, Twitter users use hashtags to add context and metadata to tweets making twitter more expressive [20]. Hashtags, indeed, provide a way to search for any kind of content posted by any kind of user, e.g. personal feelings, public criticism, nonsense messages, important updates [11]. Based on these, it is possible to create hashtag networks in Twitter by converting the co-occurrences of these tags in a single tweet into links. In several cases, these types of networks exhibit the same universal properties as real networks, such as the small-world property, clustering and power-law distributions [11, 21]. The majority of research in this context focuses on the dissemination of information and the analysis of the networks of interactions formed among users around different topics, such as popular television broadcasts [11], health care [14], COVID-19 pandemics [10]. This kind of study relates to the SNA that govern the relationships between NFTs, their creators and their collectors as described in [23] where the auction dynamics of NFTs, links between artists and collectors and co-bidding networks were studied, concluding strong evidence of first-mover advantage. Other studies were conducted using data obtained from DLT or marketplaces. In the context of NFTs transactions, the study of how different wallets interact was performed in [7] In this study, an attempt was made to identify the wallets that most influence trades using graph-based analysis techniques. Preliminary studies between the relationship between the average price of tokens in an NFT collection and twitter trends was analysed in [17] and [18]. Both works applied learning techniques to predict the future price of assets. Although the datasets considered are in both cases limited and restricted to specific NFT collections, it is possible to conclude that social activity may be a very important feature to predict over time the collection price.