Paper: arXiv 2409.00843

Authors: Yuqi Chen, Yifan Li, Kyrie Zhixuan Zhou, Xiaokang Fu, Lingbo Liu, Shuming Bao, Daniel Sui, Luyao Zhang

Abstract

Blockchain technology and decentralized finance (DeFi) are reshaping global financial systems. Despite their impact, the spatial distribution of public sentiment and its economic and geopolitical determinants are often overlooked. This study analyzes over 150 million geo-tagged, DeFi-related tweets from 2012 to 2022, sourced from a larger dataset of 7.4 billion tweets. Using sentiment scores from a BERT-based multilingual classification model, we integrated these tweets with economic and geopolitical data to create a multimodal dataset. Employing techniques like sentiment analysis, spatial econometrics, clustering, and topic modeling, we uncovered significant global variations in DeFi engagement and sentiment. Our findings indicate that economic development significantly influences DeFi engagement, particularly after 2015. Geographically weighted regression analysis revealed GDP per capita as a key predictor of DeFi tweet proportions, with its impact growing following major increases in cryptocurrency values such as bitcoin. While wealthier nations are more actively engaged in DeFi discourse, the lowest-income countries often discuss DeFi in terms of financial security and sudden wealth. Conversely, middle-income countries relate DeFi to social and religious themes, whereas high-income countries view it mainly as a speculative instrument or entertainment. This research advances interdisciplinary studies in computational social science and finance and supports open science by making our dataset and code available on GitHub, and providing a non-code workflow on the KNIME platform. These contributions enable a broad range of scholars to explore DeFi adoption and sentiment, aiding policymakers, regulators, and developers in promoting financial inclusion and responsible DeFi engagement globally.

Complexity vs Empirical Score

  • Math Complexity: 4.0/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Street Traders — empirical and applied, lower math complexity

Why this score: The study uses spatial econometrics and regression, which involves some mathematical complexity, but it is not excessively dense or advanced. The empirical rigor is high due to the analysis of over 150 million geo-tagged tweets, integration with economic data, use of BERT-based sentiment analysis, and public availability of datasets and code.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Spatiotemporal Analysis of Global<br>DeFi Sentiment & Determinants"] --> B["Data Collection & Processing<br>150M geo-tagged tweets, 2012-2022<br>+ Economic/Geopolitical Data"]
  B --> C["Core Methodology<br>Sentiment Analysis (BERT)<br>Spatial Econometrics & Clustering<br>Topic Modeling"]
  C --> D["Key Finding 1<br>DeFi Engagement driven by<br>Economic Development & GDP"]
  C --> E["Key Finding 2<br>Sentiment Themes differ by Income Level<br>Speculation (High) vs Security (Low)"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

We provide a multimodal dataset of geo‑tagged DeFi tweets with sentiment scores and open‑source code, enabling scholars to explore DeFi adoption and sentiment globally.

Source sentence
This research advances interdisciplinary studies in computational social science and finance and supports open science by making our dataset and code available on GitHub, and providing a non-code workflow on the KNIME platform.

Assumes

No explicit assumptions found in the text.

Data

  • Harvard University Twitter Archive v2.0 · Harvard University text-news public
    2010-01 to 2023-06 · global geo‑tagged tweets
    Source sentence
    DATA A. Geotagged Tweets In this study, we utilized the Harvard University Twitter Archive v2.0 [36], which comprises geotagged, multilingual tweets from around the world, spanning the period from January 2010 to June 2023.
  • Twitter Sentiment Geographical Index (TSGI) Dataset · Harvard University text-news public
    2012-2022 · 163 countries/regions
    Source sentence
    B. Twitter Sentiment Geographical Index We obtained the sentiment scores of geotagged tweets from the Twitter Sentiment Geographical Index (TSGI) Dataset [38] (https://gis.harvard.edu/twitter-sentiment-geographical-index-tsgi-dataset-global-high-frequency-dataset-monitoring).
  • World Development Indicators (GDP per capita) · World Bank macro public
    2012-2022 · countries
    Source sentence
    C. World Development Indicators We utilized data from the World Bank’s World Development Indicators (WDI), specifically focusing on the GDP per capita metric. Considering that our Twitter data spans from 2012, we calculated the average GDP per capita for different countries from 2012 to 2022, providing an accurate representation of each country’s economic status during this decade and allowing us to analyze how
  • Cryptocurrency Regulation Tracker · Atlantic Council other public
    60 countries
    Source sentence
    D. Cryptocurrency Regulation We incorporated the data from ‘Cryptocurrency Reg-ulation Tracker’ (https://www.atlanticcouncil.org/programs/ geoeconomics-center/cryptoregulationtracker/), which focuses on the regulatory landscape of cryptocurrencies across 60 countries.

Recognised sources: Twitter / X

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Economic development significantly influences DeFi engagement, particularly after 2015. statistical
    Source sentence
    Our findings indicate that economic development significantly influences DeFi engagement, particularly after 2015.
  • GDP per capita is a key predictor of the proportion of DeFi‑related tweets, with its impact growing following major cryptocurrency price increases. statistical
    Source sentence
    Geographically weighted regression analysis revealed GDP per capita as a key predictor of DeFi tweet proportions, with its impact growing following major increases in cryptocurrency values such as bitcoin.
  • Wealthier nations are more actively engaged in DeFi discourse; low‑income countries discuss DeFi in terms of financial security, middle‑income countries link it to social and religious themes, and high‑income countries view it mainly as speculative or entertai statistical
    Source sentence
    While wealthier nations are more actively engaged in DeFi discourse, the lowest-income countries often discuss DeFi in terms of financial security and sudden wealth. Conversely, middle-income countries relate DeFi to social and religious themes, whereas high-income countries view it mainly as a speculative instrument or entertainment.
  • Pseudo R2 increases over the years, indicating a growing relationship between GDP per capita and DeFi discussions, with the strongest coefficients between 2016 and 2018. statistical figure from elsewhere in the paper
    Source sentence
    The increase in R2 over the years suggests a growing relationship between GDP per capita and decentralized finance discussions.Key Observations: Before 2015, GDP per capita had an inconsistent effect on the proportion of decentralized finance-related tweets, with some years showing negative or insignificant relationships; Starting in 2015, the effect of GDP per capita became statistically significant, indicating
  • The spatial lag coefficient strengthens in later years, suggesting that regional spillover effects in DeFi discussions become more pronounced over time. statistical
    Source sentence
    indicating that economic development increasingly influenced decentralized finance discussions; The highest coefficients occur between 2016 and 2018, suggesting a peak period where wealthier countries saw greater public engagement in decentralized finance discussions on Twitter; The spatial lag coefficient (λ) grows stronger in later years, suggesting that regional spillover effects in decentralized finance

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 16 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers in the archive

Code

https://github.com/yukiyuqichen/GeoBlockchain — the authors' own repository

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (82)

  1. Corporate culture: Evidence from the field (2022)
  2. Modernization, Cultural Change, and Democracy: The Human Development Sequence (2005)
  3. AI governance in the public sector: Three tales from the frontiers of automated decision-making in democratic settings (2020)
  4. What determines AI adoption in companies? Mixed-method evidence (2024)
  5. Impact of artificial intelligence investment on firm value (2022)
  6. Trust in Traditional Finance and Consumer Fintech Adoption (2024)
  7. How are texts analyzed in blockchain research? A systematic literature review (2024)
  8. A survey of blockchain from the perspectives of applications, challenges, and opportunities (2019)
and 74 more
  1. Decentralized finance (2020)
  2. DeFi and the Future of Finance (2021) — link
  3. The design principle of blockchain: An initiative for the sok of soks (2023) — link
  4. Twitter Sentiment Geographical Index Dataset (2023) — link
  5. Harvard CGA Geotweet Archive v2.0 (2023) — link
  6. An Anatomy of Cryptocurrency Sentiment (2023) — link
  7. On the Mechanics of NFT Valuation: AI Ethics and Social Media (2023) — link
  8. “Centralized or Decentralized?”: Concerns and Value Judgments of Stakeholders in the Non-Fungible Tokens (NFTs) Market — → 2311.10990
  9. Frequency Distribution of the Values of the Correlation Coefficient in Samples from an Indefinitely Large Population (1915)
  10. Decentralized finance research and developments around the world (2022)
  11. Untangling blockchain: A data processing view of blockchain systems (2018)
  12. Blockchain-based database in an IoT environment: challenges, opportunities, and analysis (2020)
  13. Blockchain for good? Digital ledger technology and sustainable development goals (2018)
  14. What happens in blockchain stays in blockchain. A legal solution to conflicts between digital ledgers and privacy rights (2020)
  15. Blockchain with internet of things: Benefits, challenges, and future directions (2018)
  16. Anticipating the economic benefits of blockchain (2017)
  17. Measuring the perceived benefits of implementing blockchain technology in the banking sector (2021)
  18. Decentralized finance: On blockchain-and smart contract-based financial markets (2021)
  19. Decentralising Finance using Decentralised Blockchain Oracles (2020) — link
  20. Decentralized finance—A systematic literature review and research directions (2022)
  21. Blockchain technology overview (2019) — link
  22. Powered by Blockchain technology, DeFi (Decentralized Finance) strives to increase financial inclusion of the unbanked by reshaping the world financial system (2021)
  23. The Global Findex Database 2017: Measuring financial inclusion and opportunities to expand access to and use of financial services (2020)
  24. Spatial aggregation: theory and applications (1996)
  25. Spatial aggregation: Data model and implementation (2009)
  26. Finding groups in data: an introduction to cluster analysis (2009)
  27. Cluster analysis for researchers (2004)
  28. Review on determining number of Cluster in K-Means Clustering (2013)
  29. Cluster quality analysis using silhouette score (2020)
  30. Social media: Defining, developing, and divining (2015)
  31. The impact of social media on society (2015)
  32. Social media its impact with positive and negative aspects (2016)
  33. The history of social media and its impact on business (2011)
  34. When can social media lead financial markets? (2014)
  35. Crypto-currencies narrated on tweets: a sentiment analysis approach (2020)
  36. The effect of tweets made by cryptocurrency opinion leaders on bitcoin prices (2020)
  37. Impact of Social Media Sentiment on Cryptocurrency Price Movements (2022)
  38. How social network sentiments and activities impact cryptocurrencies’ trading volume and volatility? (2022)
  39. Spatial autocorrelation (2009)
  40. An improved spatial error model for the mass appraisal of commercial real estate based on spatial analysis: Shenzhen as a case study (2015)
  41. Spatial autoregressive with a spatial autoregressive error term model and its parameter estimation with two-stage generalized spatial least square procedure (2019)
  42. Fisher transformation based confidence intervals of correlations in fixed-and random-effects meta-analysis (2022)
  43. Topic modeling: a comprehensive review (2019)
  44. A bibliometric analysis of topic modelling studies (2000–2017) (2021)
  45. The exact distribution of Moran’s I (1995)
  46. Spatial processes: models & applications (1981)
  47. Spatial analysis: a guide for ecologists (2014)
  48. The interpretation of statistical maps Journal of the Royal Statistical Society Series B (Methodological) 10 (2): 243-251 (1948)
  49. Latent Dirichlet allocation (LDA) and topic modeling: models, applications, a survey (2019)
  50. Consortium blockchain-enabled smart ESG reporting platform with token-based incentives for corporate crowdsensing (2022)
  51. Blockchain-enabled ESG reporting framework for sustainable supply chain (2021)
  52. Decoding social sentiment in dao: A comparative analysis of blockchain governance communities (2023) — link
  53. Sentiment analysis in the era of large language models: A reality check. ArXiv (2023) — link
  54. Extracting Cryptocurrency Price Movements from the Reddit Network Sentiment (2019) — link
  55. DAM: A Universal Dual Attention Mechanism for Multimodal Timeseries Cryptocurrency Trend Forecasting — → 2405.00522
  56. How corruption contributes to poverty (2014)
  57. Causality between corruption, poverty and growth: a panel data analysis (2005)
  58. Corruption and poverty: A review of recent literature (2003)
  59. Inter-temporal changes in well-being during conditions of hyperinflation: evidence from Zimbabwe (2014)
  60. Poverty reduction without economic growth? explaining Brazil’s poverty dynamics, 1985-2004 (1985)
  61. Cryptocurrency and digital assets: A positive tool for economic growth in developing countries (2022)
  62. Understanding demographic and socioeconomic biases of geotagged Twitter users at the county level (2019)
  63. Population bias in geotagged tweets (2015)
  64. Bitcoin mining: A global review of energy and power demand (2019)
  65. Policy assessments for the carbon emission flows and sustainability of Bitcoin blockchain operation in China (2021)
  66. Climate concerns and the future of nonfungible tokens: Leveraging environmental benefits of the Ethereum Merge (2023)
  67. Climate sustainability through a dynamic duo: Green hydrogen and crypto driving energy transition and decarbonization (2024)
  68. Twitter Users by Country 2023 (2023)
  69. Cryptocurrencies and future financial crime (2022)
  70. The rise in popularity of cryptocurrency and associated criminal activity (2020)
  71. Convenience theory of cryptocurrency crime: A content analysis of US federal court decisions (2021)
  72. Cryptocurrency: A tool and target for cybercrime (2018)
  73. Assessing reliability of Chinese geotagged social media data for spatiotemporal representation of human mobility (2022)
  74. Twitter Sentiment Geographical Index Dataset (2023)

2 of 82 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.