⛓️🪂 Digital Methods for Blockchain Research

Exploring the Social Media Vernacular in Crypto-Finance with a Focus on Airdrops

Team Members

Janna Joceli Omena, Johnnatan Messias, Alberto Cossu, Fabio Gouveia, Augusto Falcão, Riccardo Ventura, Andrea Benedetti, Minye Huang, Linqi Ye

Links: ˚˚ Poster ˚ Relevant Crypto Airdrops for Making Lists of Keywords & Digital Objects ˚ Methods of Query Design for Airdrops Datasets ˚ List of Crypto Airdrops and Datasheet4datasets

Contents

Findings in a nutshell

Telegram

→ Visual Content: Predominantly features screenshots detailing token allocations, airdrop instructions, and eligibility criteria. 📸📋

→ Discussions: Generally professional and technical, often involving code reviews. 💻🔍

→ Community Culture: Collaborative, with members openly sharing experiences and tutorials. 🤗✨

YouTube

→ Content Style: Videos often mix real people with artificial backgrounds, using imagery that conveys money and urgency. 💵⏰

→ Demographics: The platform has a male-dominated presence, with many thumbnails featuring AI-generated figures. 👨‍💻🧑‍🎤

→ Engagement Trends: Content engagement peaks around airdrop announcements, lasting 10 days for long-term and retroactive airdrops, and 2-3 days for short-term ones. ⏳🚀

1. Introduction

Crypto-finance is a deeply social technology. It is quintessentially digital as its assets are intangible and because it relies extensively on digitally mediated communicative practices. Departing from a counter-institutional moment that stemmed from the 2008 financial crisis, at present is an industry valued at $2.55 trillion(Forbes, 2024), with an estimated 500 million investors and thousands of crypto-assets being exchanged on a variety of platforms and crypto-exchanges. Crypto-assets are also highly differentiated, including those that are now mainstream, as bitcoin and ethereum, with alternative more decentralised projects (DeFi) gaining popularity. One notable development is also the existence of crypto-coins (having their own native blockchain) and the crypto-tokens, which depend on other blockchains (the most notable one being Ethereum).

In this context, the study of Airdrop farming presents an opportunity to dive deeper into the latest trends in how crypto-finance is a realm of social interaction that can be studied via digital methods, to gain insights into both the digital cultures they create and the economic imaginaries they enact.

Historically, Airdrops can be understood as a new way for crypto-enabled projects to gain social and financial traction. Compared with Initial Coin Offerings (ICO), a precedent form of financing cryptos in 2016-2018, This project will assess if and to what extent we can attest to a change in the reward structures and the promises that the crypto-industry is marketing. ICOs were marketed (and heavily promoted by crypto-influencers) as opportunities to ‘get rich quick’, with promises to multiply by 100x the initial investment in a matter of weeks. A utopian mode often ridden with scam practices. In the case of Airdrops a tentative hypothesis is that we might be facing a more grounded, modest and delayed reward structure based on the execution of a number of promotional activities.

In a nutshell: from getting rich quick from your investment, to getting paid, maybe, for your work. Different dimensions concur with the formation of this phenomenon, including the cultural zeitgeist (the belief in technology to disintermediate and deliver more efficient financial returns, the role of online communities, etc.). What remains central for the analysis and the two-week project is:

• The relevance of technological mediations (as in the case of the web3 browser extensions) as mediums through which users enter and manage their assets

• The relevance of the social and cultural domains as they are expressed by the relevance that digital communications have in creating value. Departing from an ontological belief of the price of assets being based on a sort of inherent value what we witness in general in the crypto-financial world is that the social work conducted by ‘productive publics’ (Arvidsson 2013) is of the utmost importance to understand their cultural and therefore economic relevance.

Method background

Computational social scientists have focused on big data approaches, such as analysing BitcoinTalk online forum data over the years – May 2014 to December 2022 (Zilius, Spiliotopoulos & Van Moorsel, 2023), and drawing correlations between Twitter engagement practices and the prices of cryptocurrencies to predict the future price values of Bitcoin (Tandon et al. 2021). In digital humanities and social sciences, blockchain research has primarily focused on conceptual frameworks, as indicated by a cross-disciplinary review of blockchain research using 2,125 articles published between 2008 and early 2019 (Shahid & Jungpil, 2020). The development of quali-quanti digital methods (Omena et al., 2024; Venturini, 2024) to study blockchain technology with and about social media and web3 environments is still under investigation. Additionally, little progress has been made in studying, measuring, and verifying to what extent social media practices have concurrently contributed to or intervened in blockchain transactions.

About this project

Crypto airdrops are reward schemes that involve sending tokens to web3 wallet addresses in exchange for building a long-term loyal community that will "generate genuine economic activity" (Messias, Yaish & Livshits, 2023). Historically, airdrops have emerged as a new method for crypto projects to gain social and financial traction, differing from the get-rich-quick schemes of Initial Coin Offerings (ICOs) prevalent in 2016-2018. This project investigates seven crypto airdrops characterized by the demand for retroactive and future actions from blockchain users in exchange for receiving tokens. By examining their reward practices through the perspectives of YouTube and Telegram content creators, user engagement, and visual content production, we mapped YouTube and Telegram airdrop vernaculars using qualitative and quantitative methods with temporal datasets.

Project aims (Week 1):

• Curate robust and diverse datasets using query design methods (Rogers, 2023) to map, describe, and visualise airdrop farming practices across social media.

• Explore the Web3 software wallets and NFT platforms marketplaces.

Project aims (Week 2):

• Examine the common practices, terminologies, and visual languages in airdrop farming discussions on YouTube and Telegram, utilising quali-quanti visual methods.

2. Initial Datasets

We focused on Telegram and YouTube because they offer accessible digital data and deep insights into crypto-finance cultures. The list of crypto airdrops and data curated were documented in a datasheet for datasets. The number of unique videos and images is detailed below. The datasets were merged based on airdrop characteristics if they demanded retroactive or future actions from blockchain users in exchange for receiving tokens.

Summary of query design methods for curating crypto airdrop venarculars on YouTube and Telegram.

822 NFT images were scraped from OpenSea top images and across chains (Arbitrum, Avalanche, Base, Blast, Ethereum, Klaytn, Optimism, Polygon, Sei, Solana, and Zora). We used Image Download Script (Gouveia, 2024) to get NFTs by collection via OpenSea API.

Screen Capture shows the blockchains on which the NFT images were scraped. The images are part of the OpenSea Top category, referring to NFTs that have some consistency in the market rather than being categorised as trends.

2.1 Situating YouTube Dataset Building

Dataset building

• Search queries by categories: airdrop itself, eligibility criteria, airdrop farming

• Parameters for API calls: 10 iterations, relevance and date, specific time (Before airdrop announcement and after airdrop release)

  • Specific regions codes were used to the ZKsync airdrop datasets: United States (US), China (CN), Switzerland (CH), Singapore (SG), India (IN), Nigeria (NG), Vietnam (VN), Ukraine (UA), Brazil (BR), South Africa (ZA), Kenya (KE), Argentina (AR), United Kingdom (GB), El Salvador (SV), France (FX), Germany (DE), Japan (JP).

• Three final datasets:

  • Retroactive crypto airdrops (2022-2024) [15,222]: Starknet, Optimism, Taiko

  • Future actions crypto airdrops (2023-2024) :

    • Short term: Arbitrum, Celestia, Polyhedra [5,934]

    • Long term: ZKsync [22,701]

Besides the fact that there is no overlap among the future actions airdrop video lists, they do overlap with retroactive actions airdrops video lists and ZKsync video list (ZKsync total video IDs is 22701). There is no overlap among the future actions airdrop video lists.

Visualising the overlap among unique video lists obtained for each airdrop and according to the datasets.

2.2 Situating Telegram Dataset Building

The list of channels and groups was curated using Teleteg Search Engine, the GPTs search engine for Telegram, and data collection with TeleCatch (Ruscica, Tucci & Carneiro, 2023). From a list of 279 Telegram groups/channels, our final dataset reflects data from 38 unique channels and two unique groups and their approximately 1000 messages between the ZKsync announcement date (10 April 2023) and release (11 June 2024). The peaks in the chart below are precisely on the date of announcement and release of this crypto airdrop, and the number of messages increases with the same trend as the number of authors.

The middle peak is on 2023-11-18, with one message from only one user. The content is about an accusation of crypto being a lie. The reference is attached below.

2.3 Situating the list of tweets

We compiled a list of official announcement tweets from crypto companies conducting airdrops to analyse user reactions. Due to time constraints, the analysis was not completed, but the list is available below. X lists of tweets and comments will integrate future research.

3. Research Questions

[Week 1] Methods of Query Design

How do we map social media airdrop vernaculars using query design methods?

What are the challenges of methods of query design to map crypto airdrop vernaculars?

[Week 2] Exploratory cross-platform analysis of crypto airdrop data and visualising OpenSea NFT image collections

What can we learn about the crypto airdrop hunter community by detecting engagement peaks over time?

What are the visual vernaculars of airdrop content creators on YouTube and Telegram?

4. Methodology

4.1 Description of query design and dataset building methods

In the first week, we curated Airdrop datasets across platforms (YouTube, Telegram, and Twitter by implementing methods of query design for curating Airdrop datasets. We listed eight crypto airdrops (see table below) from L2 layer, considering their position on the DappRadar ranking. They focus on scalability (descentralization). We gathered information about the airdrops by searching for the name of the Airdrop, including the company that issued it, its announcement and release time, the name of the token issued by it, the method of Airdrop distribution, and the user's eligibility requirements. This process (documented here) provided starting points for a navigational procedure where we tested and defined search queries on YouTube and Telegram.

The process of defining YouTube search queries

YouTube search terms were initially classified into five categories (see below). After testing multiple combinations for each category and considering the recommended videos on the platform, three final categories led the data curation process: airdrop itself, eligibility criteria and airdrop farmers and sybil attacks.

• Airdrop itself (token typology, official announcement, white papers, projects that create credibility to attract users)

  • E.g. ZKsync airdrop, $ZKS airdrop, #ZKsyncairdrop

• Eligibility criteria (Selection criteria to earn the airdrop tasks, level of difficulty level, users, contributor)

  • E.g. Sort by relevance, date upload and rating: eligibility criteria ZKsync airdrop, ZKsync Airdrop Breakdown, ZKsync Airdrop guide, How to be eligible for ZKsync airdrop, How to Qualify for ZKsync Airdrop, ZKsync Airdrop claim, ZKsync airdrop rewards

• Distribution (e.g. ZKsync - 17.5% of the overall supply will be distributed through a one-time airdrop)

• Reward (types, e.g. ZKsync - Varies based on points earned and activity on ZKsyn; Optimism - Based on governance delegation, gas rebates, and additional attributes)

• Airdrop farmers and Sybil attacks

  • E.g. creating wallets for ZKsync airdrop, how to farm ZKsync airdrop, farming ZKsync airdrop, telegram bot ZKsync airdrop

The keywords were determined to be airdrop itself, eligibility criteria and airdrop farming related, of which Celestia's airdrop release was mainly focused on project promotion, so the search was conducted using the project keyword " ". In addition, since both ZKsync and Polyhendra use "ZK" as the abbreviation for token, the name of the company issuing the token should be added to the search.

4.2 Quali-quanti visual methods

We advanced quantitative analysis of the crypto airdrop hunter community by examining unique actos (e.g., unique channel IDs) and detecting engagement peaks over time. Such analysis was combined with a qualitative approach to visual content that considered the image composition, language in use (and video format for YouTube video thumbnails). Data visualisation is a device for inquiry and raising new questions.

4.3 Visualising OpenSea NFT images

To visualise the OpenSea NFT images, we created a layered image montage using ImageJ and its macro image montage feature. Although a comparative analysis of cross-blockchain NFT images was started during the data sprint, we have decided to include it in our future work. The images are available in this document for public scrutiny or for colleagues interested in analysing them.

5. Findings

The following findings are based on the process of dataset curation and three key datasets: retroactive crypto airdrops (2022–2024) from Starknet, Optimism, and Taiko [15,222 unique videos, 13,287 images]; future actions airdrops (2023–2024) for short-term projects like Arbitrum, Celestia, and Polyhedra [5,934 unique videos, 7,709 images]; and long-term airdrops for ZKsync [22,701 unique videos, 19,692 images]. Telegram data from 38 unique channels and two unique groups and their approximately 1000 messages between the ZKsync announcement date (10 April 2023) and release (11 June 2024).

5.1 On methods of query design to map crypto airdrop on social media

Choosing search queries requires integrating a good understanding of the study topic and the platform search fields and recommendation systems. This requires an iterative process and avoiding YouTube personal recommendations and algorithm-driven results. For example, when searching for "Celestia airdrop", the recommended videos were irrelevant and not directly associated with the airdrop. After verifying the technical documentation about the project, we have learnt that this crypto airdrop was addressed as "the genesis drop". When using the latter keyword, the results were relevant videos. Another example is avoiding being distracted by YouTube 's personalised recommendations when searching for videos and new potential keywords. Ignoring such recommendations and scrolling down the page has helped us identify new keywords.

Additionally, understanding crypto airdrop reward structures provided insights into creating search query categories and future data exploration strategies. For example, crypto-publics can be divided into those motivated by financial gains and those by technological interests, a distinction observable through different airdrop reward structures. Future work will explore this division in the dataset.

5.2 Quantifying YouTube Analysis: Retroactive and Future Airdrops

Retroactive crypto airdrops

YouTube Video Content Characteristics

The themes of the retroactive airdrops focus on eligibility and claiming guides.

The engagement peaks align with airdrop announcements and releases and it lasts ~10 days.

The number of videos uploaded was 15,414 videos.

Channel Characteristics

There are about 5000 unique creators active on YouTube over time.

There are 3 creators who lead in uploads and engagement.

The channels are primarily crypto-related, not just airdrops.

Future actions crypto airdrops

YouTube Video Content Characteristics

The themes of the short-term airdrops focus on eligibility and claiming guides.

The engagement peaks align with airdrop announcements and releases and it lasts ~3 days.

The number of videos uploaded was 17,600 videos.

Channel Characteristics

There are about 5000 unique creators active on YouTube overtime.

There are 10 creators who lead in uploads and engagement.

The channels are primarily crypto-related, not just airdrops.

YouTube channel characteristics

The number of creators and videos share the same pattern over time, with about 5000 unique creators detected for each type of airdrop mentioned above. About 5 to 10 unique and dominant creators were detected based on the number of videos uploaded and the engagement gained within the timeframe of analysis. The content of the channels is mainly crypto-specialised, not airdrop specific.

5.3 ZKsync Airdrop unchained with quali-quanti visual methods

ZKsync airdrop YouTube Content and Channel Characteristics

• The themes of the video content for long-term airdrop (Ksync) focus on eligibility and claiming guides.

• The engagement peaks align with airdrop announcements and releases and the peaks last ~10 days.

• There were 22,701 unique uploaded videos.

• There were ~5000 unique creators with 10 creators leading in uploads and engagement.

• The channels are primarily crypto-related, not just airdrops.

ZKsync Airdrop YouTube Image Analysis

Summary findings:

• Increase in Telegram Bot Advertising: Telegram bot advertising content has increased, possibly due to ZKsync's long-term announcement period.

• Cross-Platform Communication: Compared to short-term airdrops, there is more content related to other social media platforms, indicating more frequent cross-platform communication.

Detailed findings:

1. Unique Aesthetics:

The vast majority of YouTube covers use a mix of real people and artificial backgrounds。The inclusion of the presenter's image in the cover enhances the personalisation of the video, and that the cover character's facial expressions and poses help to create an emotional connection with the viewer as well as increase click-through rates.

2. Money Focus:

In content creation, money symbols such as dollar signs and gold coin icons are often used to grab the attention of viewers. These symbols convey the potential promise of getting rich quick and can be effective in generating viewer interest. In addition, creators often emphasise low-barrier entry conditions, such as free access or the promise of ‘get rich quick and share in the business’, in order to appeal to a wider audience. This exploits the ‘zero price effect’ in behavioural economics (Shampanier, Mazar, & Ariely, 2007, p. 743).

3. Decentralised and Sharing Community:

In many online communities, there is no single dominant channel, but rather a decentralised, interconnected approach to sharing knowledge and experience. This ecology encourages every community member to share. In addition, creators often take a step-by-step approach to sharing operational knowledge to help community members progress quickly.

4. Male-Dominated Imagery:

Many content covers feature predominantly male images, reflecting traditional gender role assignments (Connell, 2005, p. 79). At the same time, some creators use AI-generated images of women to increase the appeal of their content, a practice that may reinforce gender stereotypes (Gill, 2007, p. 101)

5. Time Emphasis:Imagery:

Creators use words like ‘soon’ or ‘now’ to create a sense of urgency and encourage viewers to take action immediately. They tend to create a sense of urgency that increases user interest and the likelihood of viewing.

ZKsync Airdrop Telegram Content Characteristics

• The peaks of views on Telegram align with airdrop announcements and releases.

• The number of messages increases with the same trend as the number of authors.

• The content of the message is about the accusation of airdrop scam.

ZKsync Airdrop Telegram images analysis

Summary findings:

• Screenshots Predominate: Most Telegram group images are screenshots highlighting key details like token allocations, airdrop instructions, and eligibility timeframes.

• Professional and Technical: Discussions are professional and technical, featuring code reviews and detailed analyses.

• Sharing Culture: Members share experiences and tutorials step-by-step.

• Close-Knit Community and MEME Culture: In close-knit groups, members express negative views more freely and use MEMEs extensively.

Detailed findings:

1. Screenshots Predominate:

Screenshots have become the primary form of communication in many Telegram groups. These screenshots often contain key information such as details of token allocations, descriptions of airdrop campaigns, and eligibility timeframes. The visual presentation of screenshots is efficient in digital communication because it conveys information near losslessly, reducing the possibility of misinterpretation of information .

2. Professional and Technical:

Another distinguishing feature of Telegram groups is their strong sharing culture. Members often share their experiences and tutorials on a step-by-step basis to help other members solve problems or improve their skills. This behaviour not only demonstrates the community's spirit of cooperation, but also facilitates the spread of knowledge and the enhancement of skills. In technical domains, knowledge sharing can significantly enhance the technical competence of members and accelerate the learning process of the community as a whole (Lave & Wenger, 1991, pp. 29-33).

3. Sharing Culture:

Another distinguishing feature of Telegram groups is their strong sharing culture. Members often share their experiences and tutorials on a step-by-step basis to help other members solve problems or improve their skills. This behaviour not only demonstrates the community's spirit of cooperation, but also facilitates the spread of knowledge and the enhancement of skills. In technical domains, knowledge sharing can significantly enhance the technical competence of members and accelerate the learning process of the community as a whole (Lave & Wenger, 1991, pp. 29-33).

4. Close-Knit Community and MEME Culture:

Members of Telegram groups are often very close-knit, and this tight social structure provides members with a relatively free space to express their thoughts, including negative opinions. The MEME culture is widely used in these groups.MEME is not only a form of humorous expression, but also a means of reinforcing a sense of community identity and belonging (Shifman, 2014, p. 72).

6. Conclusions

The study of online cultures of crypto-finance is premised on its quintessential digital nature of both the assets which are circulated and the bulk of the communication which takes place on platforms. This domain is a vast ecosystem of knowledge and emotions presenting a complex interplay between end-users (retail investors), crypto-influencers (the content creators) and technical systems (blockchain architectures). In our research we are able to follow closely the digital traces of content creators (their activities on airdrop communication, their visuals) and by combining them with our document analysis on the several airdrops we have considered (analysing their reward structure), we were able to infer the relationship it articulates between them and the publics who populate the crypto-scene.

Following this path, we believe it’s productive to analytically separate crypto-publics into two main categories: those who are (drawn) “in for the money” and those who are in (stay) “for the technology”. Such division is the result of the operationalisation by studying airdrops that exhibit two different reward structures: the first is based on rewarding a user for past actions, and the second is based on future actions that a user knows she has to perform. Future work will seek evidence of this division within our dataset to study the digital cultures of finance through airdrops.

Regarding query design methods, the selection of words and digital objects as entry points for data curation should not solely depend on the topic under study (crypto airdrops) or the specific web environment being investigated (social media), as these cannot be separated. Because we thoroughly investigated the list of airdrops—such as by asking about the companies behind them, identifying official resources, and determining the dates of announcement and release—we have gained insights into specific airdrop reward structures. These insights influenced our decision-making in creating search query categories and provided us with guidance on how we might want to explore our data. Additionally, an iterative process that embraces web environments from a methodological standpoint is crucial. This approach necessitates, for example, disregarding personal recommendations and algorithm-driven top results, as these are inherent features of social media platforms.

7. References

Arvidsson, A. (2013). The potential of consumer publics. Ephemera. Retrieved from https://ephemerajournal.org/contribution/potential-consumer-publics

Casino et al., 2019. A systematic literature review of blockchain-based applications: Current status, classification, and open issues. Telematics and Informatics, 36 (2019), pp. 55-81. https://doi.org/10.1016/j.tele.2018.11.006

“Cryptocurrency Prices, Market Cap and Charts.” 2024. Forbes. Forbes Magazine. Accessed June 14. https://www.forbes.com/digital-assets/crypto-prices/.

Messias, J., Yaish, A., & Livshits, B. (2023). Airdrops: Giving Money Away Is Harder Than It Seems. ArXiv. /abs/2312.02752

Connell, R. W. (2005). Masculinities (p. 79). University of California Press.

Gill, R. (2007). Gender and the Media (p. 101). Polity Press.

Gouveia, F.C. (2024). DownloadImages (script), GitHub repository, https://github.com/fgouveia/downloadimages

Lave, J. (1991). Situated learning: Legitimate peripheral participation. The Press Syndicate of the University of Cambridge.

Omena, J. J., Lobo, T., Tucci, G., Bitencourt, E., de Keulenaar, E., Kerche, F., Chao, J., Liedtke, M., Li, M., Paschoal, M. L., & Lavrov, I. (2024). Quali-quanti visual methods and political bots: A cross-platform study of pro- & anti-bolsobots. Journal of Digital Social Research, 6(1), 50-73. https://doi.org/10.33621/jdsr.v6i1.215

Shampanier, K., Mazar, N., & Ariely, D. (2007). Zero as a special price: The true value of free products. Marketing Science, 26(6), 742-757.

Tandon, C., Revankar, S., Palivela, H., & Parihar, S. S. (2021). How can we predict the impact of the social media messages on the value of cryptocurrency? insights from big data analytics. International Journal of Information Management Data Insights, 1(2), 100035. https://doi.org/10.1016/j.jjimei.2021.100035

Rogers R. (2023). Doing digital methods (2nd edition). Sage.

Ruscica, G., Tucci, G., & Carneiro, B. (2023). TeleCatch. GitLab. https://gitlab.com/LaBaffa/teledash

Shahid, M.N. & Jungpil, H. (2020) A Cross-Disciplinary Review of Blockchain Research Trends and Methodologies: Topic Modeling Approach. ​​Proceedings of the 53rd Hawaii International Conference on System Sciences. DOI:10.24251/HICSS.2020.495

Shifman, L. (2014). Memes in Digital Culture (p. 72). MIT Press.

Venturini, T. (2024). Quali-quantitave methods. In A. Irwin & U. Felt (Eds.), Encyclopedia of Science and Technology Studies (forthcoming).

Wang, Q., Li, R., Wang, Q., Chen, S., Ryan, M., & Hardjono, T. (2022). Exploring Web3 From the View of Blockchain. ArXiv. /abs/2206.08821

Zilius, K., Spiliotopoulos, T., & Van Moorsel, A. (2023). A dataset of coordinated cryptocurrency-related social media campaigns. Proceedings of the International AAAI Conference on Web and Social Media, 17, 1112–1121. https://doi.org/10.1609/icwsm.v17i1.22219