Методы наук о данных в политических исследованиях: анализ протестной активности в социальных сетях

Появление и рост популярности социальных сетей, а также растущая цифровизация, проникающая в разнообразные сферы экономики и общества, оказали существенное влияние на сферу политики в целом и на процессы политической мобилизации и коммуникации в частности. Методологический арсенал политической науки также оказался затронут указанными трансформационными процессами и начал пополняться новыми подходами и методами, предложенными в рамках недавно возникшей области знания, получившей название наук о данных. В статье предложен обзор ключевых инноваций в методологии исследований политической мобилизации в социальных сетях, которые были заимствованы из области наук о данных. Авторы подробно рассматривают методы обучения с учителем и без учителя и обсуждают их применимость для автоматизированного сбора данных в почти реальном времени и анализа собранных данных о протестной активности. В контексте методов обучения с учителем особое внимание уделяется методам преодоления переобучения с помощью регуляризации и выбору гиперпараметров с помощью кросс-валидации. В рамках обучения без учителя рассматриваются методы тематического моделирования и методы анализа социальных сетей. Преимущества и недостатки обсуждаемых методов иллюстрируются примерами из современных политических исследований, опубликованных в ведущих рецензируемых журналах. В заключение обсуждаются новейшие методные разработки наук о данных, до сих пор не получившие своего применения в исследованиях политической мобилизации, обладающие высоким аналитическим потенциалом (включая методы с частичным обучением, использование машинного обучения для каузального анализа и использование векторного представления текстов).

1. Ахременко А.С., Стукал Д.К., Петров А.П. Сеть или текст? Факторы распространения протеста в социальных медиа: теория и анализ данных // Полис. Политические исследования. 2020. № 2. C. 73-91. 10.17976/ jpps/2020.02.06. DOI: 10.17976/jpps/2020.02.06 EDN: APZWMB

2. Иберла К. Факторный анализ / пер. с нем. В.М. Ивановой. М.: Статистика, 1980. 398 с.

3. Azar E. The conflict and peace data bank (COPDAB) project // Journal of Conflict Resolution. 1980. Vol. 24, N 1. P. 143-152. 10.1177/ 002200278002400106. DOI: 10.1177/002200278002400106

4. Basilevsky A. Statistical factor analysis and related methods. N.Y.: Wiley, 1994. 759 p.

5. Big data, social media, and protest: foundations for a research agenda // Computational social science: discovery and prediction / J. Tucker [et al.]; M. Alvarez (ed.). N.Y.: Cambridge University Press, 2016. P. 199-224.

6. Blei D., Lafferty J. A correlated topic model of science // Annals of Applied Statistics. 2007. Vol. 1, N 1. P. 17-35. DOI: 10.1214/07-aoas114

7. Blei D., Ng A., Jordan M. Latent Dirichlet Allocation // Journal of Machine Learning Research. 2003. Vol. 3, N 3. P. 993-1022.

8. Boschee E., Natarajan P., Weischedel R. Automatic extraction of events from open source text for predictive forecasting // Handbook of computational approaches to counterterrorism / V. Subrahmanian (ed). N.Y.: Springer, 2013. P. 51-67.

9. Casas A., Williams N. Images that matter: online protests and the mobilizing role of pictures // Political Research Quarterly. 2019. Vol. 72, N 2. P. 360-375. DOI: 10.1177/1065912918786805

10. Cioffi-Revilla C. Computational social science // Wiley interdisciplinary reviews: Computational statistics. 2010. Vol. 2, N 3. P. 259-271. DOI: 10.1002/wics.95

11. Clarke K., Kocak K. Launching revolution: social media and the Egyptian uprising's first movers // British Journal of Political Science. 2020. Vol. 50, N 3. P. 1025-1045. DOI: 10.1017/s0007123418000194

12. Cluster analysis / B. Everitt, S. Landau, M. Leese, D. Stahl. Chichester: Wiley, 2011. 330 p.

13. Comparing GDELT and ICEWS Event Data / M. Ward, A. Beger, J. Cutler, [et al]. 2013. 10 p. Mode of access: https://www.researchgate.net/profile/Andreas_Beger2/publication/303211430_Comparing_GDELT_and_ICEWS_event_ data/links/57f7d9bb08ae886b89836115/Comparing-GDELT-and-ICEWS-event-data.pdf (accessed: 19.10.2020).

14. Computational social Science / D. Lazer, A. Pentland, L. Adamic, [et al.] // Science. 2009. Vol. 323, N 5915. P. 721-723. DOI: 10.1126/science.1167742

15. Croicu M., Weidmann N. Improving the selection of news reports for event coding using ensemble classification // Research and Politics. 2015. Vol. 2, N 4. P. 1-8. DOI: 10.1177/2053168015615596

16. Diamond L. Liberation technology // Journal of Democracy. 2010. Vol. 21, N 3. P. 69-83. DOI: 10.1353/jod.0.0190

17. Distributed representations of words and phrases and their compositionality / T. Mikolov, I. Sutskever, K. Chen, G. Corrado, J. Dean // Proceedings of the 26 th International Conference on Neural Information Processing Systems. Neural Information Processing Systems Foundation, 2013. P. 3111-3119. Mode of access: https://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf (accessed: 19.10.2020).

18. Elites Tweet to get feet off the streets: measuring regime social media strategies during protest / K. Munger, R. Bonneau, J. Nagler, J. Tucker // Political science research and methods. 2019. Vol. 7, N 4. P. 815-834. DOI: 10.1017/psrm.2018.3

19. Emirbayer M., Goodwin J. Network analysis, culture, and the problem of agency // American journal of sociology 1994. Vol. 99, N 6. P. 1411-1454. DOI: 10.1086/230450 EDN: BTLYCR

20. Enikolopov R., Makarin A., Petrova M. Social media and protest participation: evidence from Russia // Econometrica. 2020. Vol. 88, N 4. P. 1479-1514. DOI: 10.3982/ecta14281 EDN: WUHFHJ

21. Evaluation methods for topic models / H. Wallach, I. Murray, R. Salakhutdinov, D. Mimno // ICML'09: Proceedings of the 26 th Annual International Conference on Machine Learning. 2009. P. 1105-1112 Mode of access: https://mimno.infosci.cornell.edu/papers/wallach09evaluation.pdf (accessed: 19.10.2020).

22. Freedman D. Statistical models and shoe leather // Sociological Methodology 1991. Vol. 21 P. 291-313. DOI: 10.2307/270939

23. Golbeck J. Analyzing the social web. Amsterdam: Morgan Kaufmann, 2013. 290 p.

24. Grandjean M. A social network analysis of Twitter: Mapping the digital humanities community // Cogent Arts & Humanities. 2016. Vol. 3, N 1. P. 1-14. DOI: 10.1080/23311983.2016.1171458

25. Griffiths T., Steyvers M. Probabilistic topic models // Latent Semantic Analysis: A Road to Meaning / E. Laurence, D. Landauer, S. McNamara, D. Kintsch (eds). Mahwah, NJ: Laurence Erlbaum, 2006. P. 427-448.

26. Hafner-Burton E., Kahler M., Montgomery A. Network analysis for international relations // International organization. 2009. Vol. 63, N 3. P. 559-592. DOI: 10.1017/s0020818309090195

27. Hastie T., Tibshirani R., Friedman J. The elements of statistical learning: data mining, inference, and prediction. N.Y.: Springer, 2016 b. 745 p. DOI: 10.1007/978-0-387-84858-7

28. Hastie T., Tibshirani R., Wainwright M. Statistical learning with sparsity: the lasso and generalizations. Boca Rato: CRC Press, 2016 a. 354 p.

29. Hierarchical Dirichlet processes / Y. Teh, M. Jordan, M. Beal, D. Blei // Journal of the American Statistical Association. 2006. Vol. 101, N 476. P. 1566-1581. DOI: 10.1198/016214506000000302

30. Hoerl E., Kennard R. Ridge regression: biased estimation for nonorthogonal problems // Technometrics. 1970. Vol. 12, N 1. P. 55-67. DOI: 10.1080/00401706.1970.10488634

31. IN-SPIRE InfoVis 2004 Contest Entry / P. Wong, E. Hetzler, S. Posse, M. Whiting, S. Havre, N. Cramer, A. Shah, M. Singhal, A. Turner, J. Thomas // IEEE Symposium on Information Visualization. InfoViz. 2004. 2 p. Mode of access: https://www.cs.umd.edu/hcil/InfovisRepository/contest-2004/3/PNNLsummary2004.pdf (accessed: 19.10.2020).

32. Integrated data for events analysis (IDEA): an event typology for automated events data development / D. Bond, J. Bond, C. Oh, J. Jenkins, C.L. Taylor // Journal of peace research. 2003. Vol. 40, N 6. P. 733-745. DOI: 10.1177/00223433030406009 EDN: JNXWTD

33. Kadushin C. Too much investment in social capital? // Social networks. 2004. Vol. 1, N 26. P. 75-90. DOI: 10.1016/j.socnet.2004.01.009

34. Kadushin C. Understanding social networks: Theories, concepts, and findings. Oxford: Oxford university press, 2012. 264 p.

35. Krueger J., Lewis-Beck M. Is OLS dead? // The political methodologist. 2008. Vol. 15, N 2. P. 2-4.

36. Lankina T., Tertytchnaya K. Protest in electoral autocracies: a new dataset // Post-Soviet affairs. 2020. Vol. 36, N 1. P. 20-36. DOI: 10.1080/1060586x.2019.1656039 EDN: NQRGRU

37. Leetaru K., Schrodt P. Gdelt: Global data on events, location, and tone, 1979-2012 // ISA annual convention. 2012. Vol. 2. P. 1-49.

38. Lindgren S. Movement mobilization in the age of hashtag activism: examining the challenge of noise, hate, and disengagement in the #MeToo campaign // Policy and Internet. 2019. Vol. 11, N 4. P. 418-438. DOI: 10.1002/poi3.212

39. McClelland C. World event/interaction survey codebook (ICPSR 5211). Ann Arbor: University consortium for political and social research, 1976. 22 p.

40. Mixed-initiative social media analytics at the World Bank. Observations of citizen sentiment in Twitter Data to explore "Trust" of political actors and state institutions and its relationship to social protest / N. Calderon, B. Fisher, J. Hemsley, B. Ceskavich, G. Jansen, R. Marciano, V. Lemieux // 2015 IEEE International Conference on Big Data (Big Data). Santa Clara, CA: IEEE, 2015. P. 1678-1687. DOI: 10.1109/BigData.2015.7363939

41. Molina M., Garip F. Machine learning for sociology // Annual review of sociology. 2019. Vol. 45. P. 27-45. DOI: 10.1146/annurev-soc-073117-041106

42. Mourão R., Chen W. Covering protests on Twitter: The influences on journalists' social media portrayals of Leftand Right-Leaning demonstrations in Brazil // The International Journal of Press/Politics. 2020. Vol. 25, N 2. P. 260-280. DOI: 10.1177/1940161219882653

43. O'Brien S. Crisis early warning and decision support: contemporary approaches and thoughts on future research // International studies review. 2010. Vol. 12, N 1. P. 87-104. DOI: 10.1111/j.1468-2486.2009.00914.x

44. Open Event Data Alliance. PLOVER: Political language ontology for verifiable event records. Event, actor and data interchange specification. 2020. Mode of access: https://github.com/openeventdata/PLOVER/blob/master/PLOVER_MANUAL.pdf (accessed: 19.10.2020).

45. Pachinko Prediction: A Bayesian method for event prediction from social media data / J. Tuke, A. Nguyen, M. Nasim, D. Mellor, A. Wickramasinghe, N. Bean, L. Mitchell // Information processing and management. 2020. Vol. 57, N 2. P. 1-13. DOI: 10.1016/j.ipm.2019.102147

46. Padjett J., Ansell C. Robust Action and the Rise of the Medici // American Journal of Sociology. 1993. Vol. 98, N 6. P. 1259-1319. 10.1086/ 230190. DOI: 10.1086/230190 EDN: BOJGAJ

47. Pan J., Siegel A. How Saudi crackdowns fail to silence online dissent // American Political science review. 2020. Vol. 114, N 1. P. 109-125. DOI: 10.1017/s0003055419000650

48. Pennington J., Socher R., Manning C. Glove: global vectors for word representation // Conference on Empirical methods in natural language processing (EMNLP). Association for computational linguistics, 2014. P. 1532-1543.

49. Predicting online protest participation of social media users / S. Ranganath, F. Morstatter, X. Hu, J. Tang, S. Wang, H. Liu // 30 th AAAI Conference on artificial intelligence, AAAI 2016. Phoenix: AAAI press, 2016. P. 208-214.

50. Reading tea leaves: how humans interpret topic models / J. Chang, J. Boyd-Graber, S. Gerrish, W. Chong, D. Blei // Advances in neural information processing systems 22 (neural information processing systems 2009) / Y. Bengio [et al.] (eds). 2009. Mode of access: https://papers.nips.cc/paper/3700-reading-tea-leaves-how-humans-interpret-topic-models.pdf (accessed: 19.10.2020).

51. Rheault L., Cochrane C. Word embeddings for the analysis of ideological placement in parliamentary corpora // Political analysis. 2020. Vol. 22, N 1. P. 112-133. DOI: 10.1017/pan.2019.26

52. Roberts M., Brandon S., Dustin T. Navigating the Local Modes of Big Data: The Case of Topic Models // Data Analytics in Social Science, Government, and Industry / M. Alvarez (ed.). N.Y.: Cambridge University Press, 2016. P. 51-97.

53. Schrodt P., Gerner D., Yilmaz O. Conflict and mediation event observations (CAMEO): an event data framework for a post Cold War world // International conflict mediation: new approaches and findings / J. Bercovitch, S. Gartner (eds). N.Y.: Routledge, 2009. P. 287-304.

54. Schrodt P., Van Brackle D. Automated Coding of Political Event Data // Handbook of Computational Approaches to Counterterrorism / V. Subrahmanian (ed.). N.Y.: Springer, 2013. P. 23-49.

55. Social networks and protest participation: evidence from 130 million Twitter users / J. Larson, J. Nagler, J. Ronen, J. Tucker // American Journal of Political Science. 2019. Vol. 63, N 3. P. 690-705. DOI: 10.1111/ajps.12436

56. Steenbergen M., Bradford J. Modeling multilevel data structures // American Journal of Political Science. 2002. Vol. 46, N 1. P. 218-237. DOI: 10.2307/3088424

57. Structural topic models for open-ended survey responses / M. Roberts, S. Brandon, T. Dustin, C. Lucas, L.-L. Jetson, S. Gadarian, B. Albertson, D. Rand // American journal of political science. 2014. Vol. 58, N 4. P. 1064-1082. DOI: 10.1111/ajps.12103

58. Sundberg R., Melander E. Introducing the UCDP georeferenced event dataset // Journal of peace research. 2013. Vol. 50, N 4. P. 523-532. DOI: 10.1177/0022343313484347

59. The structural topic model and applied social science / M. Roberts, S. Brandon, T. Dustin, E. Airoldi // Advances in neural information processing systems workshop on topic models: computation, application, and evaluation. NIPS, 2013. Mode of access: https://scholar.princeton.edu/files/bstewart/files/stmnips2013.pdf (accessed: 19.10.2020).

60. Tibshirani R. Regression Shrinkage and selection via the Lasso // Journal of the Royal statistical society. Series B (methodological). 1996. Vol. 58, N 1. P. 267-288. DOI: 10.1111/j.2517-6161.1996.tb02080.x

61. Tremayne M. Anatomy of protest in the digital era: a network analysis of Twitter and Occupy Wall Street // Social movement studies. 2014. Vol. 13, N 1. P. 110-126. DOI: 10.1080/14742837.2013.830969 EDN: YDXUAT

62. Tweeting from left to right: Is online political communication more than an echo chamber? / P. Barberá, J.T. Jost, J. Nagler, J.A. Tucker, R. Bonneau // Psychological science. 2015. Vol. 26, N 10. P. 1531-1542. 10.1177/ 0956797615594620. DOI: 10.1177/0956797615594620

63. Vera E., Schupp T. Network analysis in comparative social sciences // Comparative Education. 2006. Vol. 42, N 3. P. 405-429. 10.1080/ 03050060600876723. DOI: 10.1080/03050060600876723

64. ViEWS: A political violence early-warning system / H. Hegre, M. Allansson, M. Basedau, [et al.] // Journal of Peace Research. 2019. Vol. 56, N 2. P. 155-174. DOI: 10.1177/0022343319823860

65. Weidmann N., Rød E. The Internet and political protest in autocracies. Chapter 4 // Weidmann N., Rød E. The Internet and Political Protest in Autocracies. N.Y.:Oxford university press, 2019. P. 39-62.

66. Wooldridge J. Econometric analysis of cross section and panel data. Cambridge: The MIT Press, 2002. 741 p.

67. Zhu X., Goldberg A. Introduction to semi-supervised learning // Synthesis lectures on artificial intelligence and machine learning. 2009. Vol. 3, N 1. P. 1-130. DOI: 10.2200/S00196ED1V01Y200906AIM006

68. Zhuravskaya E., Petrova M., Enikolopov R. Political effects of the internet and social media // Annual review of economics. 2020. Vol. 12. P. 415-438. DOI: 10.1146/annurev-economics-081919-050239 EDN: PECTYV